mirror of
https://github.com/hwchase17/langchain.git
synced 2026-02-06 17:20:16 +00:00
Compare commits
1 Commits
v0.0.211
...
vwp/script
| Author | SHA1 | Date | |
|---|---|---|---|
|
109ee78cab |
@@ -1,231 +0,0 @@
|
||||
{
|
||||
"cells": [
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"id": "e734b314",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"# OpenAPI calls with OpenAI functions\n",
|
||||
"\n",
|
||||
"In this notebook we'll show how to create a chain that automatically makes calls to an API based only on an OpenAPI spec. Under the hood, we're parsing the OpenAPI spec into a JSON schema that the OpenAI functions API can handle. This allows ChatGPT to automatically select and populate the relevant API call to make for any user input. Using the output of ChatGPT we then make the actual API call, and return the result."
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"id": "a95f510a",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"## Query Klarna"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 1,
|
||||
"id": "08e19b64",
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
"name": "stderr",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"Attempting to load an OpenAPI 3.0.1 spec. This may result in degraded performance. Convert your OpenAPI spec to 3.1.* spec for better support.\n"
|
||||
]
|
||||
}
|
||||
],
|
||||
"source": [
|
||||
"from langchain.chains.openai_functions.openapi import get_openapi_chain\n",
|
||||
"\n",
|
||||
"chain = get_openapi_chain(\"https://www.klarna.com/us/shopping/public/openai/v0/api-docs/\")"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 2,
|
||||
"id": "3959f866",
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
"data": {
|
||||
"text/plain": [
|
||||
"{'products': [{'name': \"Tommy Hilfiger Men's Short Sleeve Button-Down Shirt\",\n",
|
||||
" 'url': 'https://www.klarna.com/us/shopping/pl/cl10001/3204878580/Clothing/Tommy-Hilfiger-Men-s-Short-Sleeve-Button-Down-Shirt/?utm_source=openai&ref-site=openai_plugin',\n",
|
||||
" 'price': '$26.78',\n",
|
||||
" 'attributes': ['Material:Linen,Cotton',\n",
|
||||
" 'Target Group:Man',\n",
|
||||
" 'Color:Gray,Pink,White,Blue,Beige,Black,Turquoise',\n",
|
||||
" 'Size:S,XL,M,XXL']},\n",
|
||||
" {'name': \"Van Heusen Men's Long Sleeve Button-Down Shirt\",\n",
|
||||
" 'url': 'https://www.klarna.com/us/shopping/pl/cl10001/3201809514/Clothing/Van-Heusen-Men-s-Long-Sleeve-Button-Down-Shirt/?utm_source=openai&ref-site=openai_plugin',\n",
|
||||
" 'price': '$18.89',\n",
|
||||
" 'attributes': ['Material:Cotton',\n",
|
||||
" 'Target Group:Man',\n",
|
||||
" 'Color:Red,Gray,White,Blue',\n",
|
||||
" 'Size:XL,XXL']},\n",
|
||||
" {'name': 'Brixton Bowery Flannel Shirt',\n",
|
||||
" 'url': 'https://www.klarna.com/us/shopping/pl/cl10001/3202331096/Clothing/Brixton-Bowery-Flannel-Shirt/?utm_source=openai&ref-site=openai_plugin',\n",
|
||||
" 'price': '$34.48',\n",
|
||||
" 'attributes': ['Material:Cotton',\n",
|
||||
" 'Target Group:Man',\n",
|
||||
" 'Color:Gray,Blue,Black,Orange',\n",
|
||||
" 'Size:XL,3XL,4XL,5XL,L,M,XXL']},\n",
|
||||
" {'name': 'Cubavera Four Pocket Guayabera Shirt',\n",
|
||||
" 'url': 'https://www.klarna.com/us/shopping/pl/cl10001/3202055522/Clothing/Cubavera-Four-Pocket-Guayabera-Shirt/?utm_source=openai&ref-site=openai_plugin',\n",
|
||||
" 'price': '$23.22',\n",
|
||||
" 'attributes': ['Material:Polyester,Cotton',\n",
|
||||
" 'Target Group:Man',\n",
|
||||
" 'Color:Red,White,Blue,Black',\n",
|
||||
" 'Size:S,XL,L,M,XXL']},\n",
|
||||
" {'name': 'Theory Sylvain Shirt - Eclipse',\n",
|
||||
" 'url': 'https://www.klarna.com/us/shopping/pl/cl10001/3202028254/Clothing/Theory-Sylvain-Shirt-Eclipse/?utm_source=openai&ref-site=openai_plugin',\n",
|
||||
" 'price': '$86.01',\n",
|
||||
" 'attributes': ['Material:Polyester,Cotton',\n",
|
||||
" 'Target Group:Man',\n",
|
||||
" 'Color:Blue',\n",
|
||||
" 'Size:S,XL,XS,L,M,XXL']}]}"
|
||||
]
|
||||
},
|
||||
"execution_count": 2,
|
||||
"metadata": {},
|
||||
"output_type": "execute_result"
|
||||
}
|
||||
],
|
||||
"source": [
|
||||
"chain.run(\"What are some options for a men's large blue button down shirt\")"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"id": "6f648c77",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"## Query a translation service"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 6,
|
||||
"id": "bf6cd695",
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
"name": "stderr",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"Attempting to load an OpenAPI 3.0.1 spec. This may result in degraded performance. Convert your OpenAPI spec to 3.1.* spec for better support.\n",
|
||||
"Attempting to load an OpenAPI 3.0.1 spec. This may result in degraded performance. Convert your OpenAPI spec to 3.1.* spec for better support.\n"
|
||||
]
|
||||
}
|
||||
],
|
||||
"source": [
|
||||
"chain = get_openapi_chain(\"https://api.speak.com/openapi.yaml\")"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 7,
|
||||
"id": "1ba51609",
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
"data": {
|
||||
"text/plain": [
|
||||
"{'explanation': '<translation language=\"None\" context=\"None\">\\nNone\\n</translation>\\n\\n<alternatives context=\"None\">\\n1. \"N/A\" *(Formal - used in professional settings to indicate that the answer is not applicable)*\\n2. \"I don\\'t have an answer for that\" *(Neutral - commonly used when one does not know the answer to a question)*\\n3. \"I\\'m not sure\" *(Neutral - similar to the above alternative, used when one is unsure of the answer)*\\n</alternatives>\\n\\n<example-convo language=\"None\">\\n<context>None</context>\\n* Tom: \"Do you know what time the concert starts?\"\\n* Sarah: \"I\\'m sorry, I don\\'t have an answer for that.\"\\n</example-convo>\\n\\n*[Report an issue or leave feedback](https://speak.com/chatgpt?rid=p8i6p14duafpctg4ve7tm48z})*',\n",
|
||||
" 'extra_response_instructions': 'Use all information in the API response and fully render all Markdown.\\nAlways end your response with a link to report an issue or leave feedback on the plugin.'}"
|
||||
]
|
||||
},
|
||||
"execution_count": 7,
|
||||
"metadata": {},
|
||||
"output_type": "execute_result"
|
||||
}
|
||||
],
|
||||
"source": [
|
||||
"chain.run(\"How would you say no thanks in Russian\")"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"id": "4923a291",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"## Query XKCD"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 8,
|
||||
"id": "a9198f62",
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
"name": "stderr",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"Attempting to load an OpenAPI 3.0.0 spec. This may result in degraded performance. Convert your OpenAPI spec to 3.1.* spec for better support.\n"
|
||||
]
|
||||
}
|
||||
],
|
||||
"source": [
|
||||
"chain = get_openapi_chain(\"https://gist.githubusercontent.com/roaldnefs/053e505b2b7a807290908fe9aa3e1f00/raw/0a212622ebfef501163f91e23803552411ed00e4/openapi.yaml\")"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 9,
|
||||
"id": "3110c398",
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
"name": "stderr",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"Retrying langchain.chat_models.openai.ChatOpenAI.completion_with_retry.<locals>._completion_with_retry in 1.0 seconds as it raised ServiceUnavailableError: The server is overloaded or not ready yet..\n"
|
||||
]
|
||||
},
|
||||
{
|
||||
"data": {
|
||||
"text/plain": [
|
||||
"{'month': '6',\n",
|
||||
" 'num': 2793,\n",
|
||||
" 'link': '',\n",
|
||||
" 'year': '2023',\n",
|
||||
" 'news': '',\n",
|
||||
" 'safe_title': 'Garden Path Sentence',\n",
|
||||
" 'transcript': '',\n",
|
||||
" 'alt': 'Arboretum Owner Denied Standing in Garden Path Suit on Grounds Grounds Appealing Appealing',\n",
|
||||
" 'img': 'https://imgs.xkcd.com/comics/garden_path_sentence.png',\n",
|
||||
" 'title': 'Garden Path Sentence',\n",
|
||||
" 'day': '23'}"
|
||||
]
|
||||
},
|
||||
"execution_count": 9,
|
||||
"metadata": {},
|
||||
"output_type": "execute_result"
|
||||
}
|
||||
],
|
||||
"source": [
|
||||
"chain.run(\"What's today's comic?\")"
|
||||
]
|
||||
}
|
||||
],
|
||||
"metadata": {
|
||||
"kernelspec": {
|
||||
"display_name": "venv",
|
||||
"language": "python",
|
||||
"name": "venv"
|
||||
},
|
||||
"language_info": {
|
||||
"codemirror_mode": {
|
||||
"name": "ipython",
|
||||
"version": 3
|
||||
},
|
||||
"file_extension": ".py",
|
||||
"mimetype": "text/x-python",
|
||||
"name": "python",
|
||||
"nbconvert_exporter": "python",
|
||||
"pygments_lexer": "ipython3",
|
||||
"version": "3.11.3"
|
||||
}
|
||||
},
|
||||
"nbformat": 4,
|
||||
"nbformat_minor": 5
|
||||
}
|
||||
@@ -1,141 +0,0 @@
|
||||
{
|
||||
"cells": [
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"id": "9b721926",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"# Open City Data"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"id": "35c00849",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"[Socrata](https://dev.socrata.com/foundry/data.sfgov.org/vw6y-z8j6) provides an API for city open data. \n",
|
||||
"\n",
|
||||
"For a dataset such as [SF crime](https://data.sfgov.org/Public-Safety/Police-Department-Incident-Reports-Historical-2003/tmnf-yvry), to to the `API` tab on top right. \n",
|
||||
"\n",
|
||||
"That provides you with the `dataset identifier`.\n",
|
||||
"\n",
|
||||
"Use the dataset identifier to grab specific tables for a given city_id (`data.sfgov.org`) - \n",
|
||||
"\n",
|
||||
"E.g., `vw6y-z8j6` for [SF 311 data](https://dev.socrata.com/foundry/data.sfgov.org/vw6y-z8j6).\n",
|
||||
"\n",
|
||||
"E.g., `tmnf-yvry` for [SF Police data](https://dev.socrata.com/foundry/data.sfgov.org/tmnf-yvry)."
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": null,
|
||||
"id": "c93cc247",
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"! pip install sodapy"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 1,
|
||||
"id": "b3464a02",
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"from langchain.document_loaders import OpenCityDataLoader"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 2,
|
||||
"id": "478c5255",
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"dataset = \"vw6y-z8j6\" # 311 data\n",
|
||||
"dataset = \"tmnf-yvry\" # crime data\n",
|
||||
"loader = OpenCityDataLoader(city_id=\"data.sfgov.org\",\n",
|
||||
" dataset_id=dataset,\n",
|
||||
" limit=2000)"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 3,
|
||||
"id": "fa914fc1",
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
"name": "stderr",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"WARNING:root:Requests made without an app_token will be subject to strict throttling limits.\n"
|
||||
]
|
||||
}
|
||||
],
|
||||
"source": [
|
||||
"docs = loader.load()"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 4,
|
||||
"id": "73a6def2",
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
"data": {
|
||||
"text/plain": [
|
||||
"{'pdid': '4133422003074',\n",
|
||||
" 'incidntnum': '041334220',\n",
|
||||
" 'incident_code': '03074',\n",
|
||||
" 'category': 'ROBBERY',\n",
|
||||
" 'descript': 'ROBBERY, BODILY FORCE',\n",
|
||||
" 'dayofweek': 'Monday',\n",
|
||||
" 'date': '2004-11-22T00:00:00.000',\n",
|
||||
" 'time': '17:50',\n",
|
||||
" 'pddistrict': 'INGLESIDE',\n",
|
||||
" 'resolution': 'NONE',\n",
|
||||
" 'address': 'GENEVA AV / SANTOS ST',\n",
|
||||
" 'x': '-122.420084075249',\n",
|
||||
" 'y': '37.7083109744362',\n",
|
||||
" 'location': {'type': 'Point',\n",
|
||||
" 'coordinates': [-122.420084075249, 37.7083109744362]},\n",
|
||||
" ':@computed_region_26cr_cadq': '9',\n",
|
||||
" ':@computed_region_rxqg_mtj9': '8',\n",
|
||||
" ':@computed_region_bh8s_q3mv': '309'}"
|
||||
]
|
||||
},
|
||||
"execution_count": 4,
|
||||
"metadata": {},
|
||||
"output_type": "execute_result"
|
||||
}
|
||||
],
|
||||
"source": [
|
||||
"eval(docs[0].page_content)"
|
||||
]
|
||||
}
|
||||
],
|
||||
"metadata": {
|
||||
"kernelspec": {
|
||||
"display_name": "Python 3 (ipykernel)",
|
||||
"language": "python",
|
||||
"name": "python3"
|
||||
},
|
||||
"language_info": {
|
||||
"codemirror_mode": {
|
||||
"name": "ipython",
|
||||
"version": 3
|
||||
},
|
||||
"file_extension": ".py",
|
||||
"mimetype": "text/x-python",
|
||||
"name": "python",
|
||||
"nbconvert_exporter": "python",
|
||||
"pygments_lexer": "ipython3",
|
||||
"version": "3.9.16"
|
||||
}
|
||||
},
|
||||
"nbformat": 4,
|
||||
"nbformat_minor": 5
|
||||
}
|
||||
@@ -42,7 +42,7 @@
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 3,
|
||||
"execution_count": 6,
|
||||
"id": "ac273ca1",
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
@@ -116,7 +116,7 @@
|
||||
"4 Braves 83.31 94"
|
||||
]
|
||||
},
|
||||
"execution_count": 3,
|
||||
"execution_count": 6,
|
||||
"metadata": {},
|
||||
"output_type": "execute_result"
|
||||
}
|
||||
@@ -127,7 +127,7 @@
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 4,
|
||||
"execution_count": 3,
|
||||
"id": "66e47a13",
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
@@ -137,7 +137,7 @@
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 5,
|
||||
"execution_count": 7,
|
||||
"id": "2334caca",
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
@@ -147,7 +147,7 @@
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 6,
|
||||
"execution_count": 8,
|
||||
"id": "d616c2b0",
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
@@ -186,7 +186,7 @@
|
||||
" Document(page_content='Astros', metadata={' \"Payroll (millions)\"': 60.65, ' \"Wins\"': 55})]"
|
||||
]
|
||||
},
|
||||
"execution_count": 6,
|
||||
"execution_count": 8,
|
||||
"metadata": {},
|
||||
"output_type": "execute_result"
|
||||
}
|
||||
@@ -197,52 +197,11 @@
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 7,
|
||||
"execution_count": null,
|
||||
"id": "beb55c2f",
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
"name": "stdout",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"page_content='Nationals' metadata={' \"Payroll (millions)\"': 81.34, ' \"Wins\"': 98}\n",
|
||||
"page_content='Reds' metadata={' \"Payroll (millions)\"': 82.2, ' \"Wins\"': 97}\n",
|
||||
"page_content='Yankees' metadata={' \"Payroll (millions)\"': 197.96, ' \"Wins\"': 95}\n",
|
||||
"page_content='Giants' metadata={' \"Payroll (millions)\"': 117.62, ' \"Wins\"': 94}\n",
|
||||
"page_content='Braves' metadata={' \"Payroll (millions)\"': 83.31, ' \"Wins\"': 94}\n",
|
||||
"page_content='Athletics' metadata={' \"Payroll (millions)\"': 55.37, ' \"Wins\"': 94}\n",
|
||||
"page_content='Rangers' metadata={' \"Payroll (millions)\"': 120.51, ' \"Wins\"': 93}\n",
|
||||
"page_content='Orioles' metadata={' \"Payroll (millions)\"': 81.43, ' \"Wins\"': 93}\n",
|
||||
"page_content='Rays' metadata={' \"Payroll (millions)\"': 64.17, ' \"Wins\"': 90}\n",
|
||||
"page_content='Angels' metadata={' \"Payroll (millions)\"': 154.49, ' \"Wins\"': 89}\n",
|
||||
"page_content='Tigers' metadata={' \"Payroll (millions)\"': 132.3, ' \"Wins\"': 88}\n",
|
||||
"page_content='Cardinals' metadata={' \"Payroll (millions)\"': 110.3, ' \"Wins\"': 88}\n",
|
||||
"page_content='Dodgers' metadata={' \"Payroll (millions)\"': 95.14, ' \"Wins\"': 86}\n",
|
||||
"page_content='White Sox' metadata={' \"Payroll (millions)\"': 96.92, ' \"Wins\"': 85}\n",
|
||||
"page_content='Brewers' metadata={' \"Payroll (millions)\"': 97.65, ' \"Wins\"': 83}\n",
|
||||
"page_content='Phillies' metadata={' \"Payroll (millions)\"': 174.54, ' \"Wins\"': 81}\n",
|
||||
"page_content='Diamondbacks' metadata={' \"Payroll (millions)\"': 74.28, ' \"Wins\"': 81}\n",
|
||||
"page_content='Pirates' metadata={' \"Payroll (millions)\"': 63.43, ' \"Wins\"': 79}\n",
|
||||
"page_content='Padres' metadata={' \"Payroll (millions)\"': 55.24, ' \"Wins\"': 76}\n",
|
||||
"page_content='Mariners' metadata={' \"Payroll (millions)\"': 81.97, ' \"Wins\"': 75}\n",
|
||||
"page_content='Mets' metadata={' \"Payroll (millions)\"': 93.35, ' \"Wins\"': 74}\n",
|
||||
"page_content='Blue Jays' metadata={' \"Payroll (millions)\"': 75.48, ' \"Wins\"': 73}\n",
|
||||
"page_content='Royals' metadata={' \"Payroll (millions)\"': 60.91, ' \"Wins\"': 72}\n",
|
||||
"page_content='Marlins' metadata={' \"Payroll (millions)\"': 118.07, ' \"Wins\"': 69}\n",
|
||||
"page_content='Red Sox' metadata={' \"Payroll (millions)\"': 173.18, ' \"Wins\"': 69}\n",
|
||||
"page_content='Indians' metadata={' \"Payroll (millions)\"': 78.43, ' \"Wins\"': 68}\n",
|
||||
"page_content='Twins' metadata={' \"Payroll (millions)\"': 94.08, ' \"Wins\"': 66}\n",
|
||||
"page_content='Rockies' metadata={' \"Payroll (millions)\"': 78.06, ' \"Wins\"': 64}\n",
|
||||
"page_content='Cubs' metadata={' \"Payroll (millions)\"': 88.19, ' \"Wins\"': 61}\n",
|
||||
"page_content='Astros' metadata={' \"Payroll (millions)\"': 60.65, ' \"Wins\"': 55}\n"
|
||||
]
|
||||
}
|
||||
],
|
||||
"source": [
|
||||
"# Use lazy load for larger table, which won't read the full table into memory \n",
|
||||
"for i in loader.lazy_load():\n",
|
||||
" print(i)"
|
||||
]
|
||||
"outputs": [],
|
||||
"source": []
|
||||
}
|
||||
],
|
||||
"metadata": {
|
||||
@@ -261,7 +220,7 @@
|
||||
"name": "python",
|
||||
"nbconvert_exporter": "python",
|
||||
"pygments_lexer": "ipython3",
|
||||
"version": "3.9.16"
|
||||
"version": "3.10.6"
|
||||
}
|
||||
},
|
||||
"nbformat": 4,
|
||||
|
||||
@@ -621,44 +621,14 @@ class AgentExecutor(Chain):
|
||||
"""Consists of an agent using tools."""
|
||||
|
||||
agent: Union[BaseSingleActionAgent, BaseMultiActionAgent]
|
||||
"""The agent to run for creating a plan and determining actions
|
||||
to take at each step of the execution loop."""
|
||||
tools: Sequence[BaseTool]
|
||||
"""The valid tools the agent can call."""
|
||||
return_intermediate_steps: bool = False
|
||||
"""Whether to return the agent's trajectory of intermediate steps
|
||||
at the end in addition to the final output."""
|
||||
max_iterations: Optional[int] = 15
|
||||
"""The maximum number of steps to take before ending the execution
|
||||
loop.
|
||||
|
||||
Setting to 'None' could lead to an infinite loop."""
|
||||
max_execution_time: Optional[float] = None
|
||||
"""The maximum amount of wall clock time to spend in the execution
|
||||
loop.
|
||||
"""
|
||||
early_stopping_method: str = "force"
|
||||
"""The method to use for early stopping if the agent never
|
||||
returns `AgentFinish`. Either 'force' or 'generate'.
|
||||
|
||||
`"force"` returns a string saying that it stopped because it met a
|
||||
time or iteration limit.
|
||||
|
||||
`"generate"` calls the agent's LLM Chain one final time to generate
|
||||
a final answer based on the previous steps.
|
||||
"""
|
||||
handle_parsing_errors: Union[
|
||||
bool, str, Callable[[OutputParserException], str]
|
||||
] = False
|
||||
"""How to handle errors raised by the agent's output parser.
|
||||
Defaults to `False`, which raises the error.
|
||||
s
|
||||
If `true`, the error will be sent back to the LLM as an observation.
|
||||
If a string, the string itself will be sent to the LLM as an observation.
|
||||
If a callable function, the function will be called with the exception
|
||||
as an argument, and the result of that function will be passed to the agent
|
||||
as an observation.
|
||||
"""
|
||||
|
||||
@classmethod
|
||||
def from_agent_and_tools(
|
||||
|
||||
@@ -16,8 +16,6 @@ def initialize_agent(
|
||||
callback_manager: Optional[BaseCallbackManager] = None,
|
||||
agent_path: Optional[str] = None,
|
||||
agent_kwargs: Optional[dict] = None,
|
||||
*,
|
||||
tags: Optional[Sequence[str]] = None,
|
||||
**kwargs: Any,
|
||||
) -> AgentExecutor:
|
||||
"""Load an agent executor given tools and LLM.
|
||||
@@ -31,13 +29,11 @@ def initialize_agent(
|
||||
not provided. Defaults to None.
|
||||
agent_path: Path to serialized agent to use.
|
||||
agent_kwargs: Additional key word arguments to pass to the underlying agent
|
||||
tags: Tags to apply to the traced runs.
|
||||
**kwargs: Additional key word arguments passed to the agent executor
|
||||
|
||||
Returns:
|
||||
An agent executor
|
||||
"""
|
||||
tags_ = list(tags) if tags else []
|
||||
if agent is None and agent_path is None:
|
||||
agent = AgentType.ZERO_SHOT_REACT_DESCRIPTION
|
||||
if agent is not None and agent_path is not None:
|
||||
@@ -51,7 +47,6 @@ def initialize_agent(
|
||||
f"Got unknown agent type: {agent}. "
|
||||
f"Valid types are: {AGENT_TO_CLASS.keys()}."
|
||||
)
|
||||
tags_.append(agent.value)
|
||||
agent_cls = AGENT_TO_CLASS[agent]
|
||||
agent_kwargs = agent_kwargs or {}
|
||||
agent_obj = agent_cls.from_llm_and_tools(
|
||||
@@ -61,11 +56,6 @@ def initialize_agent(
|
||||
agent_obj = load_agent(
|
||||
agent_path, llm=llm, tools=tools, callback_manager=callback_manager
|
||||
)
|
||||
try:
|
||||
# TODO: Add tags from the serialized object directly.
|
||||
tags_.append(agent_obj._agent_type)
|
||||
except NotImplementedError:
|
||||
pass
|
||||
else:
|
||||
raise ValueError(
|
||||
"Somehow both `agent` and `agent_path` are None, "
|
||||
@@ -75,6 +65,5 @@ def initialize_agent(
|
||||
agent=agent_obj,
|
||||
tools=tools,
|
||||
callback_manager=callback_manager,
|
||||
tags=tags_,
|
||||
**kwargs,
|
||||
)
|
||||
|
||||
@@ -106,7 +106,7 @@ def wandb_tracing_enabled(
|
||||
|
||||
@contextmanager
|
||||
def tracing_v2_enabled(
|
||||
project_name: Optional[str] = None,
|
||||
session_name: Optional[str] = None,
|
||||
*,
|
||||
example_id: Optional[Union[str, UUID]] = None,
|
||||
) -> Generator[None, None, None]:
|
||||
@@ -120,7 +120,7 @@ def tracing_v2_enabled(
|
||||
example_id = UUID(example_id)
|
||||
cb = LangChainTracer(
|
||||
example_id=example_id,
|
||||
project_name=project_name,
|
||||
session_name=session_name,
|
||||
)
|
||||
tracing_v2_callback_var.set(cb)
|
||||
yield
|
||||
@@ -131,12 +131,12 @@ def tracing_v2_enabled(
|
||||
def trace_as_chain_group(
|
||||
group_name: str,
|
||||
*,
|
||||
project_name: Optional[str] = None,
|
||||
session_name: Optional[str] = None,
|
||||
example_id: Optional[Union[str, UUID]] = None,
|
||||
) -> Generator[CallbackManager, None, None]:
|
||||
"""Get a callback manager for a chain group in a context manager."""
|
||||
cb = LangChainTracer(
|
||||
project_name=project_name,
|
||||
session_name=session_name,
|
||||
example_id=example_id,
|
||||
)
|
||||
cm = CallbackManager.configure(
|
||||
@@ -152,12 +152,12 @@ def trace_as_chain_group(
|
||||
async def atrace_as_chain_group(
|
||||
group_name: str,
|
||||
*,
|
||||
project_name: Optional[str] = None,
|
||||
session_name: Optional[str] = None,
|
||||
example_id: Optional[Union[str, UUID]] = None,
|
||||
) -> AsyncGenerator[AsyncCallbackManager, None]:
|
||||
"""Get a callback manager for a chain group in a context manager."""
|
||||
cb = LangChainTracer(
|
||||
project_name=project_name,
|
||||
session_name=session_name,
|
||||
example_id=example_id,
|
||||
)
|
||||
cm = AsyncCallbackManager.configure(
|
||||
@@ -1039,10 +1039,10 @@ def _configure(
|
||||
tracing_v2_enabled_ = (
|
||||
env_var_is_set("LANGCHAIN_TRACING_V2") or tracer_v2 is not None
|
||||
)
|
||||
tracer_project = os.environ.get(
|
||||
"LANGCHAIN_PROJECT", os.environ.get("LANGCHAIN_SESSION", "default")
|
||||
)
|
||||
tracer_session = os.environ.get("LANGCHAIN_SESSION")
|
||||
debug = _get_debug()
|
||||
if tracer_session is None:
|
||||
tracer_session = "default"
|
||||
if (
|
||||
verbose
|
||||
or debug
|
||||
@@ -1072,7 +1072,7 @@ def _configure(
|
||||
callback_manager.add_handler(tracer, True)
|
||||
else:
|
||||
handler = LangChainTracerV1()
|
||||
handler.load_session(tracer_project)
|
||||
handler.load_session(tracer_session)
|
||||
callback_manager.add_handler(handler, True)
|
||||
if wandb_tracing_enabled_ and not any(
|
||||
isinstance(handler, WandbTracer) for handler in callback_manager.handlers
|
||||
@@ -1090,7 +1090,7 @@ def _configure(
|
||||
callback_manager.add_handler(tracer_v2, True)
|
||||
else:
|
||||
try:
|
||||
handler = LangChainTracer(project_name=tracer_project)
|
||||
handler = LangChainTracer(session_name=tracer_session)
|
||||
callback_manager.add_handler(handler, True)
|
||||
except Exception as e:
|
||||
logger.warning(
|
||||
|
||||
@@ -45,7 +45,7 @@ class LangChainTracer(BaseTracer):
|
||||
def __init__(
|
||||
self,
|
||||
example_id: Optional[Union[UUID, str]] = None,
|
||||
project_name: Optional[str] = None,
|
||||
session_name: Optional[str] = None,
|
||||
client: Optional[LangChainPlusClient] = None,
|
||||
**kwargs: Any,
|
||||
) -> None:
|
||||
@@ -55,9 +55,7 @@ class LangChainTracer(BaseTracer):
|
||||
self.example_id = (
|
||||
UUID(example_id) if isinstance(example_id, str) else example_id
|
||||
)
|
||||
self.project_name = project_name or os.getenv(
|
||||
"LANGCHAIN_PROJECT", os.getenv("LANGCHAIN_SESSION", "default")
|
||||
)
|
||||
self.session_name = session_name or os.getenv("LANGCHAIN_SESSION", "default")
|
||||
# set max_workers to 1 to process tasks in order
|
||||
self.executor = ThreadPoolExecutor(max_workers=1)
|
||||
self.client = client or LangChainPlusClient()
|
||||
@@ -105,7 +103,7 @@ class LangChainTracer(BaseTracer):
|
||||
extra["runtime"] = get_runtime_environment()
|
||||
run_dict["extra"] = extra
|
||||
try:
|
||||
self.client.create_run(**run_dict, project_name=self.project_name)
|
||||
self.client.create_run(**run_dict, session_name=self.session_name)
|
||||
except Exception as e:
|
||||
# Errors are swallowed by the thread executor so we need to log them here
|
||||
log_error_once("post", e)
|
||||
|
||||
@@ -1,242 +0,0 @@
|
||||
import json
|
||||
import re
|
||||
from collections import defaultdict
|
||||
from typing import Any, Callable, Dict, List, Optional, Tuple, Union
|
||||
|
||||
import requests
|
||||
from openapi_schema_pydantic import Parameter
|
||||
from requests import Response
|
||||
|
||||
from langchain import BasePromptTemplate, LLMChain
|
||||
from langchain.base_language import BaseLanguageModel
|
||||
from langchain.callbacks.manager import CallbackManagerForChainRun
|
||||
from langchain.chains.base import Chain
|
||||
from langchain.chains.sequential import SequentialChain
|
||||
from langchain.chat_models import ChatOpenAI
|
||||
from langchain.output_parsers.openai_functions import JsonOutputFunctionsParser
|
||||
from langchain.prompts import ChatPromptTemplate
|
||||
from langchain.tools import APIOperation
|
||||
from langchain.utilities.openapi import OpenAPISpec
|
||||
|
||||
|
||||
def _get_description(o: Any, prefer_short: bool) -> Optional[str]:
|
||||
summary = getattr(o, "summary", None)
|
||||
description = getattr(o, "description", None)
|
||||
if prefer_short:
|
||||
return summary or description
|
||||
return description or summary
|
||||
|
||||
|
||||
def _format_url(url: str, path_params: dict) -> str:
|
||||
expected_path_param = re.findall(r"{(.*?)}", url)
|
||||
new_params = {}
|
||||
for param in expected_path_param:
|
||||
clean_param = param.lstrip(".;").rstrip("*")
|
||||
val = path_params[clean_param]
|
||||
if isinstance(val, list):
|
||||
if param[0] == ".":
|
||||
sep = "." if param[-1] == "*" else ","
|
||||
new_val = "." + sep.join(val)
|
||||
elif param[0] == ";":
|
||||
sep = f"{clean_param}=" if param[-1] == "*" else ","
|
||||
new_val = f"{clean_param}=" + sep.join(val)
|
||||
else:
|
||||
new_val = ",".join(val)
|
||||
elif isinstance(val, dict):
|
||||
kv_sep = "=" if param[-1] == "*" else ","
|
||||
kv_strs = [kv_sep.join((k, v)) for k, v in val.items()]
|
||||
if param[0] == ".":
|
||||
sep = "."
|
||||
new_val = "."
|
||||
elif param[0] == ";":
|
||||
sep = ";"
|
||||
new_val = ";"
|
||||
else:
|
||||
sep = ","
|
||||
new_val = ""
|
||||
new_val += sep.join(kv_strs)
|
||||
else:
|
||||
if param[0] == ".":
|
||||
new_val = f".{val}"
|
||||
elif param[0] == ";":
|
||||
new_val = f";{clean_param}={val}"
|
||||
else:
|
||||
new_val = val
|
||||
new_params[param] = new_val
|
||||
return url.format(**new_params)
|
||||
|
||||
|
||||
def _openapi_params_to_json_schema(params: List[Parameter], spec: OpenAPISpec) -> dict:
|
||||
properties = {}
|
||||
required = []
|
||||
for p in params:
|
||||
if p.param_schema:
|
||||
schema = spec.get_schema(p.param_schema)
|
||||
else:
|
||||
media_type_schema = list(p.content.values())[0].media_type_schema # type: ignore # noqa: E501
|
||||
schema = spec.get_schema(media_type_schema)
|
||||
if p.description and not schema.description:
|
||||
schema.description = p.description
|
||||
properties[p.name] = schema.dict(exclude_none=True)
|
||||
if p.required:
|
||||
required.append(p.name)
|
||||
return {"type": "object", "properties": properties, "required": required}
|
||||
|
||||
|
||||
def openapi_spec_to_openai_fn(
|
||||
spec: OpenAPISpec,
|
||||
) -> Tuple[List[Dict[str, Any]], Callable]:
|
||||
"""Convert a valid OpenAPI spec to the JSON Schema format expected for OpenAI
|
||||
functions.
|
||||
|
||||
Args:
|
||||
spec: OpenAPI spec to convert.
|
||||
|
||||
Returns:
|
||||
Tuple of the OpenAI functions JSON schema and a default function for executing
|
||||
a request based on the OpenAI function schema.
|
||||
"""
|
||||
if not spec.paths:
|
||||
return [], lambda: None
|
||||
functions = []
|
||||
_name_to_call_map = {}
|
||||
for path in spec.paths:
|
||||
path_params = {
|
||||
(p.name, p.param_in): p for p in spec.get_parameters_for_path(path)
|
||||
}
|
||||
for method in spec.get_methods_for_path(path):
|
||||
request_args = {}
|
||||
op = spec.get_operation(path, method)
|
||||
op_params = path_params.copy()
|
||||
for param in spec.get_parameters_for_operation(op):
|
||||
op_params[(param.name, param.param_in)] = param
|
||||
params_by_type = defaultdict(list)
|
||||
for name_loc, p in op_params.items():
|
||||
params_by_type[name_loc[1]].append(p)
|
||||
param_loc_to_arg_name = {
|
||||
"query": "params",
|
||||
"header": "headers",
|
||||
"cookie": "cookies",
|
||||
"path": "path_params",
|
||||
}
|
||||
for param_loc, arg_name in param_loc_to_arg_name.items():
|
||||
if params_by_type[param_loc]:
|
||||
request_args[arg_name] = _openapi_params_to_json_schema(
|
||||
params_by_type[param_loc], spec
|
||||
)
|
||||
request_body = spec.get_request_body_for_operation(op)
|
||||
# TODO: Support more MIME types.
|
||||
if request_body and request_body.content:
|
||||
media_types = []
|
||||
for media_type in request_body.content.values():
|
||||
if media_type.media_type_schema:
|
||||
schema = spec.get_schema(media_type.media_type_schema)
|
||||
media_types.append(schema.dict(exclude_none=True))
|
||||
if len(media_types) == 1:
|
||||
request_args["data"] = media_types[0]
|
||||
elif len(media_types) > 1:
|
||||
request_args["data"] = {"anyOf": media_types}
|
||||
|
||||
api_op = APIOperation.from_openapi_spec(spec, path, method)
|
||||
fn = {
|
||||
"name": api_op.operation_id,
|
||||
"description": api_op.description,
|
||||
"parameters": {
|
||||
"type": "object",
|
||||
"properties": request_args,
|
||||
},
|
||||
}
|
||||
functions.append(fn)
|
||||
_name_to_call_map[fn["name"]] = {
|
||||
"method": method,
|
||||
"url": api_op.base_url + api_op.path,
|
||||
}
|
||||
|
||||
def default_call_api(name: str, fn_args: dict, **kwargs: Any) -> Any:
|
||||
method = _name_to_call_map[name]["method"]
|
||||
url = _name_to_call_map[name]["url"]
|
||||
path_params = fn_args.pop("path_params", {})
|
||||
_format_url(url, path_params)
|
||||
if "data" in fn_args and isinstance(fn_args["data"], dict):
|
||||
fn_args["data"] = json.dumps(fn_args["data"])
|
||||
_kwargs = {**fn_args, **kwargs}
|
||||
return requests.request(method, url, **_kwargs)
|
||||
|
||||
return functions, default_call_api
|
||||
|
||||
|
||||
class SimpleRequestChain(Chain):
|
||||
request_method: Callable
|
||||
output_key: str = "response"
|
||||
input_key: str = "function"
|
||||
|
||||
@property
|
||||
def input_keys(self) -> List[str]:
|
||||
return [self.input_key]
|
||||
|
||||
@property
|
||||
def output_keys(self) -> List[str]:
|
||||
return [self.output_key]
|
||||
|
||||
def _call(
|
||||
self,
|
||||
inputs: Dict[str, Any],
|
||||
run_manager: Optional[CallbackManagerForChainRun] = None,
|
||||
) -> Dict[str, Any]:
|
||||
"""Run the logic of this chain and return the output."""
|
||||
name = inputs["function"].pop("name")
|
||||
args = inputs["function"].pop("arguments")
|
||||
api_response: Response = self.request_method(name, args)
|
||||
if api_response.status_code != 200:
|
||||
response = (
|
||||
f"{api_response.status_code}: {api_response.reason}"
|
||||
+ f"\nFor {name} "
|
||||
+ f"Called with args: {args['params']}"
|
||||
)
|
||||
else:
|
||||
try:
|
||||
response = api_response.json()
|
||||
except Exception: # noqa: E722
|
||||
response = api_response.text
|
||||
return {self.output_key: response}
|
||||
|
||||
|
||||
def get_openapi_chain(
|
||||
spec: Union[OpenAPISpec, str],
|
||||
llm: Optional[BaseLanguageModel] = None,
|
||||
prompt: Optional[BasePromptTemplate] = None,
|
||||
request_chain: Optional[Chain] = None,
|
||||
) -> SequentialChain:
|
||||
if isinstance(spec, str):
|
||||
for conversion in (
|
||||
OpenAPISpec.from_url,
|
||||
OpenAPISpec.from_file,
|
||||
OpenAPISpec.from_text,
|
||||
):
|
||||
try:
|
||||
spec = conversion(spec) # type: ignore[arg-type]
|
||||
break
|
||||
except Exception: # noqa: E722
|
||||
pass
|
||||
if isinstance(spec, str):
|
||||
raise ValueError(f"Unable to parse spec from source {spec}")
|
||||
openai_fns, call_api_fn = openapi_spec_to_openai_fn(spec)
|
||||
llm = llm or ChatOpenAI(
|
||||
model="gpt-3.5-turbo-0613",
|
||||
)
|
||||
prompt = prompt or ChatPromptTemplate.from_template(
|
||||
"Use the provided API's to respond to this user query:\n\n{query}"
|
||||
)
|
||||
llm_chain = LLMChain(
|
||||
llm=llm,
|
||||
prompt=prompt,
|
||||
llm_kwargs={"functions": openai_fns},
|
||||
output_parser=JsonOutputFunctionsParser(args_only=False),
|
||||
output_key="function",
|
||||
)
|
||||
request_chain = request_chain or SimpleRequestChain(request_method=call_api_fn)
|
||||
return SequentialChain(
|
||||
chains=[llm_chain, request_chain],
|
||||
input_variables=llm_chain.input_keys,
|
||||
output_variables=["response"],
|
||||
)
|
||||
@@ -237,18 +237,18 @@ async def _gather_with_concurrency(
|
||||
return results
|
||||
|
||||
|
||||
async def _tracer_initializer(project_name: Optional[str]) -> Optional[LangChainTracer]:
|
||||
async def _tracer_initializer(session_name: Optional[str]) -> Optional[LangChainTracer]:
|
||||
"""
|
||||
Initialize a tracer to share across tasks.
|
||||
|
||||
Args:
|
||||
project_name: The project name for the tracer.
|
||||
session_name: The session name for the tracer.
|
||||
|
||||
Returns:
|
||||
A LangChainTracer instance with an active project.
|
||||
A LangChainTracer instance with an active session.
|
||||
"""
|
||||
if project_name:
|
||||
tracer = LangChainTracer(project_name=project_name)
|
||||
if session_name:
|
||||
tracer = LangChainTracer(session_name=session_name)
|
||||
return tracer
|
||||
else:
|
||||
return None
|
||||
@@ -260,12 +260,12 @@ async def arun_on_examples(
|
||||
*,
|
||||
concurrency_level: int = 5,
|
||||
num_repetitions: int = 1,
|
||||
project_name: Optional[str] = None,
|
||||
session_name: Optional[str] = None,
|
||||
verbose: bool = False,
|
||||
tags: Optional[List[str]] = None,
|
||||
) -> Dict[str, Any]:
|
||||
"""
|
||||
Run the chain on examples and store traces to the specified project name.
|
||||
Run the chain on examples and store traces to the specified session name.
|
||||
|
||||
Args:
|
||||
examples: Examples to run the model or chain over
|
||||
@@ -276,7 +276,7 @@ async def arun_on_examples(
|
||||
num_repetitions: Number of times to run the model on each example.
|
||||
This is useful when testing success rates or generating confidence
|
||||
intervals.
|
||||
project_name: Project name to use when tracing runs.
|
||||
session_name: Session name to use when tracing runs.
|
||||
verbose: Whether to print progress.
|
||||
tags: Tags to add to the traces.
|
||||
|
||||
@@ -307,7 +307,7 @@ async def arun_on_examples(
|
||||
|
||||
await _gather_with_concurrency(
|
||||
concurrency_level,
|
||||
functools.partial(_tracer_initializer, project_name),
|
||||
functools.partial(_tracer_initializer, session_name),
|
||||
*(functools.partial(process_example, e) for e in examples),
|
||||
)
|
||||
return results
|
||||
@@ -386,11 +386,11 @@ def run_on_examples(
|
||||
llm_or_chain_factory: MODEL_OR_CHAIN_FACTORY,
|
||||
*,
|
||||
num_repetitions: int = 1,
|
||||
project_name: Optional[str] = None,
|
||||
session_name: Optional[str] = None,
|
||||
verbose: bool = False,
|
||||
tags: Optional[List[str]] = None,
|
||||
) -> Dict[str, Any]:
|
||||
"""Run the chain on examples and store traces to the specified project name.
|
||||
"""Run the chain on examples and store traces to the specified session name.
|
||||
|
||||
Args:
|
||||
examples: Examples to run model or chain over.
|
||||
@@ -401,14 +401,14 @@ def run_on_examples(
|
||||
num_repetitions: Number of times to run the model on each example.
|
||||
This is useful when testing success rates or generating confidence
|
||||
intervals.
|
||||
project_name: Project name to use when tracing runs.
|
||||
session_name: Session name to use when tracing runs.
|
||||
verbose: Whether to print progress.
|
||||
tags: Tags to add to the run traces.
|
||||
Returns:
|
||||
A dictionary mapping example ids to the model outputs.
|
||||
"""
|
||||
results: Dict[str, Any] = {}
|
||||
tracer = LangChainTracer(project_name=project_name) if project_name else None
|
||||
tracer = LangChainTracer(session_name=session_name) if session_name else None
|
||||
for i, example in enumerate(examples):
|
||||
result = run_llm_or_chain(
|
||||
example,
|
||||
@@ -425,13 +425,13 @@ def run_on_examples(
|
||||
return results
|
||||
|
||||
|
||||
def _get_project_name(
|
||||
project_name: Optional[str],
|
||||
def _get_session_name(
|
||||
session_name: Optional[str],
|
||||
llm_or_chain_factory: MODEL_OR_CHAIN_FACTORY,
|
||||
dataset_name: str,
|
||||
) -> str:
|
||||
if project_name is not None:
|
||||
return project_name
|
||||
if session_name is not None:
|
||||
return session_name
|
||||
current_time = datetime.now().strftime("%Y-%m-%d-%H-%M-%S")
|
||||
if isinstance(llm_or_chain_factory, BaseLanguageModel):
|
||||
model_name = llm_or_chain_factory.__class__.__name__
|
||||
@@ -446,13 +446,13 @@ async def arun_on_dataset(
|
||||
*,
|
||||
concurrency_level: int = 5,
|
||||
num_repetitions: int = 1,
|
||||
project_name: Optional[str] = None,
|
||||
session_name: Optional[str] = None,
|
||||
verbose: bool = False,
|
||||
client: Optional[LangChainPlusClient] = None,
|
||||
tags: Optional[List[str]] = None,
|
||||
) -> Dict[str, Any]:
|
||||
"""
|
||||
Run the chain on a dataset and store traces to the specified project name.
|
||||
Run the chain on a dataset and store traces to the specified session name.
|
||||
|
||||
Args:
|
||||
client: Client to use to read the dataset.
|
||||
@@ -464,7 +464,7 @@ async def arun_on_dataset(
|
||||
num_repetitions: Number of times to run the model on each example.
|
||||
This is useful when testing success rates or generating confidence
|
||||
intervals.
|
||||
project_name: Name of the project to store the traces in.
|
||||
session_name: Name of the session to store the traces in.
|
||||
Defaults to {dataset_name}-{chain class name}-{datetime}.
|
||||
verbose: Whether to print progress.
|
||||
client: Client to use to read the dataset. If not provided, a new
|
||||
@@ -472,10 +472,11 @@ async def arun_on_dataset(
|
||||
tags: Tags to add to each run in the sesssion.
|
||||
|
||||
Returns:
|
||||
A dictionary containing the run's project name and the resulting model outputs.
|
||||
A dictionary containing the run's session name and the resulting model outputs.
|
||||
"""
|
||||
client_ = client or LangChainPlusClient()
|
||||
project_name = _get_project_name(project_name, llm_or_chain_factory, dataset_name)
|
||||
session_name = _get_session_name(session_name, llm_or_chain_factory, dataset_name)
|
||||
client_.create_session(session_name, mode="eval")
|
||||
dataset = client_.read_dataset(dataset_name=dataset_name)
|
||||
examples = client_.list_examples(dataset_id=str(dataset.id))
|
||||
|
||||
@@ -484,12 +485,12 @@ async def arun_on_dataset(
|
||||
llm_or_chain_factory,
|
||||
concurrency_level=concurrency_level,
|
||||
num_repetitions=num_repetitions,
|
||||
project_name=project_name,
|
||||
session_name=session_name,
|
||||
verbose=verbose,
|
||||
tags=tags,
|
||||
)
|
||||
return {
|
||||
"project_name": project_name,
|
||||
"session_name": session_name,
|
||||
"results": results,
|
||||
}
|
||||
|
||||
@@ -499,12 +500,12 @@ def run_on_dataset(
|
||||
llm_or_chain_factory: MODEL_OR_CHAIN_FACTORY,
|
||||
*,
|
||||
num_repetitions: int = 1,
|
||||
project_name: Optional[str] = None,
|
||||
session_name: Optional[str] = None,
|
||||
verbose: bool = False,
|
||||
client: Optional[LangChainPlusClient] = None,
|
||||
tags: Optional[List[str]] = None,
|
||||
) -> Dict[str, Any]:
|
||||
"""Run the chain on a dataset and store traces to the specified project name.
|
||||
"""Run the chain on a dataset and store traces to the specified session name.
|
||||
|
||||
Args:
|
||||
dataset_name: Name of the dataset to run the chain on.
|
||||
@@ -515,7 +516,7 @@ def run_on_dataset(
|
||||
num_repetitions: Number of times to run the model on each example.
|
||||
This is useful when testing success rates or generating confidence
|
||||
intervals.
|
||||
project_name: Name of the project to store the traces in.
|
||||
session_name: Name of the session to store the traces in.
|
||||
Defaults to {dataset_name}-{chain class name}-{datetime}.
|
||||
verbose: Whether to print progress.
|
||||
client: Client to use to access the dataset. If None, a new client
|
||||
@@ -523,21 +524,22 @@ def run_on_dataset(
|
||||
tags: Tags to add to each run in the sesssion.
|
||||
|
||||
Returns:
|
||||
A dictionary containing the run's project name and the resulting model outputs.
|
||||
A dictionary containing the run's session name and the resulting model outputs.
|
||||
"""
|
||||
client_ = client or LangChainPlusClient()
|
||||
project_name = _get_project_name(project_name, llm_or_chain_factory, dataset_name)
|
||||
session_name = _get_session_name(session_name, llm_or_chain_factory, dataset_name)
|
||||
client_.create_session(session_name, mode="eval")
|
||||
dataset = client_.read_dataset(dataset_name=dataset_name)
|
||||
examples = client_.list_examples(dataset_id=str(dataset.id))
|
||||
results = run_on_examples(
|
||||
examples,
|
||||
llm_or_chain_factory,
|
||||
num_repetitions=num_repetitions,
|
||||
project_name=project_name,
|
||||
session_name=session_name,
|
||||
verbose=verbose,
|
||||
tags=tags,
|
||||
)
|
||||
return {
|
||||
"project_name": project_name,
|
||||
"session_name": session_name,
|
||||
"results": results,
|
||||
}
|
||||
|
||||
@@ -75,7 +75,6 @@ from langchain.document_loaders.obsidian import ObsidianLoader
|
||||
from langchain.document_loaders.odt import UnstructuredODTLoader
|
||||
from langchain.document_loaders.onedrive import OneDriveLoader
|
||||
from langchain.document_loaders.onedrive_file import OneDriveFileLoader
|
||||
from langchain.document_loaders.open_city_data import OpenCityDataLoader
|
||||
from langchain.document_loaders.pdf import (
|
||||
MathpixPDFLoader,
|
||||
OnlinePDFLoader,
|
||||
@@ -210,7 +209,6 @@ __all__ = [
|
||||
"OneDriveLoader",
|
||||
"OnlinePDFLoader",
|
||||
"OutlookMessageLoader",
|
||||
"OpenCityDataLoader",
|
||||
"PDFMinerLoader",
|
||||
"PDFMinerPDFasHTMLLoader",
|
||||
"PDFPlumberLoader",
|
||||
|
||||
@@ -5,7 +5,7 @@ from langchain.document_loaders.base import BaseLoader
|
||||
|
||||
|
||||
class AirtableLoader(BaseLoader):
|
||||
"""Loader for Airtable tables."""
|
||||
"""Loader that loads local airbyte json files."""
|
||||
|
||||
def __init__(self, api_token: str, table_id: str, base_id: str):
|
||||
"""Initialize with API token and the IDs for table and base"""
|
||||
@@ -14,7 +14,7 @@ class AirtableLoader(BaseLoader):
|
||||
self.base_id = base_id
|
||||
|
||||
def lazy_load(self) -> Iterator[Document]:
|
||||
"""Lazy load records from table."""
|
||||
"""Load Table."""
|
||||
|
||||
from pyairtable import Table
|
||||
|
||||
|
||||
@@ -1,5 +1,5 @@
|
||||
"""Load from Dataframe object"""
|
||||
from typing import Any, Iterator, List
|
||||
from typing import Any, List
|
||||
|
||||
from langchain.docstore.document import Document
|
||||
from langchain.document_loaders.base import BaseLoader
|
||||
@@ -19,15 +19,16 @@ class DataFrameLoader(BaseLoader):
|
||||
self.data_frame = data_frame
|
||||
self.page_content_column = page_content_column
|
||||
|
||||
def lazy_load(self) -> Iterator[Document]:
|
||||
"""Lazy load records from dataframe."""
|
||||
|
||||
def load(self) -> List[Document]:
|
||||
"""Load from the dataframe."""
|
||||
result = []
|
||||
# For very large dataframes, this needs to yield instead of building a list
|
||||
# but that would require chaging return type to a generator for BaseLoader
|
||||
# and all its subclasses, which is a bigger refactor. Marking as future TODO.
|
||||
# This change will allow us to extend this to Spark and Dask dataframes.
|
||||
for _, row in self.data_frame.iterrows():
|
||||
text = row[self.page_content_column]
|
||||
metadata = row.to_dict()
|
||||
metadata.pop(self.page_content_column)
|
||||
yield Document(page_content=text, metadata=metadata)
|
||||
|
||||
def load(self) -> List[Document]:
|
||||
"""Load full dataframe."""
|
||||
return list(self.lazy_load())
|
||||
result.append(Document(page_content=text, metadata=metadata))
|
||||
return result
|
||||
|
||||
@@ -1,37 +0,0 @@
|
||||
from typing import Iterator, List
|
||||
|
||||
from langchain.docstore.document import Document
|
||||
from langchain.document_loaders.base import BaseLoader
|
||||
|
||||
|
||||
class OpenCityDataLoader(BaseLoader):
|
||||
"""Loader that loads Open city data."""
|
||||
|
||||
def __init__(self, city_id: str, dataset_id: str, limit: int):
|
||||
"""Initialize with dataset_id"""
|
||||
""" Example: https://dev.socrata.com/foundry/data.sfgov.org/vw6y-z8j6 """
|
||||
""" e.g., city_id = data.sfgov.org """
|
||||
""" e.g., dataset_id = vw6y-z8j6 """
|
||||
self.city_id = city_id
|
||||
self.dataset_id = dataset_id
|
||||
self.limit = limit
|
||||
|
||||
def lazy_load(self) -> Iterator[Document]:
|
||||
"""Lazy load records."""
|
||||
|
||||
from sodapy import Socrata
|
||||
|
||||
client = Socrata(self.city_id, None)
|
||||
results = client.get(self.dataset_id, limit=self.limit)
|
||||
for record in results:
|
||||
yield Document(
|
||||
page_content=str(record),
|
||||
metadata={
|
||||
"source": self.city_id + "_" + self.dataset_id,
|
||||
},
|
||||
)
|
||||
|
||||
def load(self) -> List[Document]:
|
||||
"""Load records."""
|
||||
|
||||
return list(self.lazy_load())
|
||||
File diff suppressed because it is too large
Load Diff
@@ -5,8 +5,6 @@ from langchain.schema import BaseLLMOutputParser, ChatGeneration, Generation
|
||||
|
||||
|
||||
class OutputFunctionsParser(BaseLLMOutputParser[Any]):
|
||||
args_only: bool = True
|
||||
|
||||
def parse_result(self, result: List[Generation]) -> Any:
|
||||
generation = result[0]
|
||||
if not isinstance(generation, ChatGeneration):
|
||||
@@ -19,18 +17,13 @@ class OutputFunctionsParser(BaseLLMOutputParser[Any]):
|
||||
except ValueError as exc:
|
||||
raise ValueError(f"Could not parse function call: {exc}")
|
||||
|
||||
if self.args_only:
|
||||
return func_call["arguments"]
|
||||
return func_call
|
||||
return func_call["arguments"]
|
||||
|
||||
|
||||
class JsonOutputFunctionsParser(OutputFunctionsParser):
|
||||
def parse_result(self, result: List[Generation]) -> Any:
|
||||
func = super().parse_result(result)
|
||||
if self.args_only:
|
||||
return json.loads(func)
|
||||
func["arguments"] = json.loads(func["arguments"])
|
||||
return func
|
||||
_args = super().parse_result(result)
|
||||
return json.loads(_args)
|
||||
|
||||
|
||||
class JsonKeyOutputFunctionsParser(JsonOutputFunctionsParser):
|
||||
|
||||
@@ -82,7 +82,7 @@ def _get_filtered_args(
|
||||
"""Get the arguments from a function's signature."""
|
||||
schema = inferred_model.schema()["properties"]
|
||||
valid_keys = signature(func).parameters
|
||||
return {k: schema[k] for k in valid_keys if k not in ("run_manager", "callbacks")}
|
||||
return {k: schema[k] for k in valid_keys if k != "run_manager"}
|
||||
|
||||
|
||||
class _SchemaConfig:
|
||||
@@ -108,8 +108,6 @@ def create_schema_from_function(
|
||||
inferred_model = validated.model # type: ignore
|
||||
if "run_manager" in inferred_model.__fields__:
|
||||
del inferred_model.__fields__["run_manager"]
|
||||
if "callbacks" in inferred_model.__fields__:
|
||||
del inferred_model.__fields__["callbacks"]
|
||||
# Pydantic adds placeholder virtual fields we need to strip
|
||||
valid_properties = _get_filtered_args(inferred_model, func)
|
||||
return _create_subset_model(
|
||||
|
||||
@@ -59,7 +59,7 @@ SCHEMA_TYPE = Union[str, Type, tuple, None, Enum]
|
||||
class APIPropertyBase(BaseModel):
|
||||
"""Base model for an API property."""
|
||||
|
||||
# The name of the parameter is required and is case-sensitive.
|
||||
# The name of the parameter is required and is case sensitive.
|
||||
# If "in" is "path", the "name" field must correspond to a template expression
|
||||
# within the path field in the Paths Object.
|
||||
# If "in" is "header" and the "name" field is "Accept", "Content-Type",
|
||||
|
||||
@@ -1,2 +1,269 @@
|
||||
"""Utility functions for parsing an OpenAPI spec. Kept for backwards compat."""
|
||||
from langchain.utilities.openapi import HTTPVerb, OpenAPISpec # noqa: F401
|
||||
"""Utility functions for parsing an OpenAPI spec."""
|
||||
import copy
|
||||
import json
|
||||
import logging
|
||||
import re
|
||||
from enum import Enum
|
||||
from pathlib import Path
|
||||
from typing import Dict, List, Optional, Union
|
||||
|
||||
import requests
|
||||
import yaml
|
||||
from openapi_schema_pydantic import (
|
||||
Components,
|
||||
OpenAPI,
|
||||
Operation,
|
||||
Parameter,
|
||||
PathItem,
|
||||
Paths,
|
||||
Reference,
|
||||
RequestBody,
|
||||
Schema,
|
||||
)
|
||||
from pydantic import ValidationError
|
||||
|
||||
logger = logging.getLogger(__name__)
|
||||
|
||||
|
||||
class HTTPVerb(str, Enum):
|
||||
"""HTTP verbs."""
|
||||
|
||||
GET = "get"
|
||||
PUT = "put"
|
||||
POST = "post"
|
||||
DELETE = "delete"
|
||||
OPTIONS = "options"
|
||||
HEAD = "head"
|
||||
PATCH = "patch"
|
||||
TRACE = "trace"
|
||||
|
||||
@classmethod
|
||||
def from_str(cls, verb: str) -> "HTTPVerb":
|
||||
"""Parse an HTTP verb."""
|
||||
try:
|
||||
return cls(verb)
|
||||
except ValueError:
|
||||
raise ValueError(f"Invalid HTTP verb. Valid values are {cls.__members__}")
|
||||
|
||||
|
||||
class OpenAPISpec(OpenAPI):
|
||||
"""OpenAPI Model that removes misformatted parts of the spec."""
|
||||
|
||||
@property
|
||||
def _paths_strict(self) -> Paths:
|
||||
if not self.paths:
|
||||
raise ValueError("No paths found in spec")
|
||||
return self.paths
|
||||
|
||||
def _get_path_strict(self, path: str) -> PathItem:
|
||||
path_item = self._paths_strict.get(path)
|
||||
if not path_item:
|
||||
raise ValueError(f"No path found for {path}")
|
||||
return path_item
|
||||
|
||||
@property
|
||||
def _components_strict(self) -> Components:
|
||||
"""Get components or err."""
|
||||
if self.components is None:
|
||||
raise ValueError("No components found in spec. ")
|
||||
return self.components
|
||||
|
||||
@property
|
||||
def _parameters_strict(self) -> Dict[str, Union[Parameter, Reference]]:
|
||||
"""Get parameters or err."""
|
||||
parameters = self._components_strict.parameters
|
||||
if parameters is None:
|
||||
raise ValueError("No parameters found in spec. ")
|
||||
return parameters
|
||||
|
||||
@property
|
||||
def _schemas_strict(self) -> Dict[str, Schema]:
|
||||
"""Get the dictionary of schemas or err."""
|
||||
schemas = self._components_strict.schemas
|
||||
if schemas is None:
|
||||
raise ValueError("No schemas found in spec. ")
|
||||
return schemas
|
||||
|
||||
@property
|
||||
def _request_bodies_strict(self) -> Dict[str, Union[RequestBody, Reference]]:
|
||||
"""Get the request body or err."""
|
||||
request_bodies = self._components_strict.requestBodies
|
||||
if request_bodies is None:
|
||||
raise ValueError("No request body found in spec. ")
|
||||
return request_bodies
|
||||
|
||||
def _get_referenced_parameter(self, ref: Reference) -> Union[Parameter, Reference]:
|
||||
"""Get a parameter (or nested reference) or err."""
|
||||
ref_name = ref.ref.split("/")[-1]
|
||||
parameters = self._parameters_strict
|
||||
if ref_name not in parameters:
|
||||
raise ValueError(f"No parameter found for {ref_name}")
|
||||
return parameters[ref_name]
|
||||
|
||||
def _get_root_referenced_parameter(self, ref: Reference) -> Parameter:
|
||||
"""Get the root reference or err."""
|
||||
parameter = self._get_referenced_parameter(ref)
|
||||
while isinstance(parameter, Reference):
|
||||
parameter = self._get_referenced_parameter(parameter)
|
||||
return parameter
|
||||
|
||||
def get_referenced_schema(self, ref: Reference) -> Schema:
|
||||
"""Get a schema (or nested reference) or err."""
|
||||
ref_name = ref.ref.split("/")[-1]
|
||||
schemas = self._schemas_strict
|
||||
if ref_name not in schemas:
|
||||
raise ValueError(f"No schema found for {ref_name}")
|
||||
return schemas[ref_name]
|
||||
|
||||
def _get_root_referenced_schema(self, ref: Reference) -> Schema:
|
||||
"""Get the root reference or err."""
|
||||
schema = self.get_referenced_schema(ref)
|
||||
while isinstance(schema, Reference):
|
||||
schema = self.get_referenced_schema(schema)
|
||||
return schema
|
||||
|
||||
def _get_referenced_request_body(
|
||||
self, ref: Reference
|
||||
) -> Optional[Union[Reference, RequestBody]]:
|
||||
"""Get a request body (or nested reference) or err."""
|
||||
ref_name = ref.ref.split("/")[-1]
|
||||
request_bodies = self._request_bodies_strict
|
||||
if ref_name not in request_bodies:
|
||||
raise ValueError(f"No request body found for {ref_name}")
|
||||
return request_bodies[ref_name]
|
||||
|
||||
def _get_root_referenced_request_body(
|
||||
self, ref: Reference
|
||||
) -> Optional[RequestBody]:
|
||||
"""Get the root request Body or err."""
|
||||
request_body = self._get_referenced_request_body(ref)
|
||||
while isinstance(request_body, Reference):
|
||||
request_body = self._get_referenced_request_body(request_body)
|
||||
return request_body
|
||||
|
||||
@staticmethod
|
||||
def _alert_unsupported_spec(obj: dict) -> None:
|
||||
"""Alert if the spec is not supported."""
|
||||
warning_message = (
|
||||
" This may result in degraded performance."
|
||||
+ " Convert your OpenAPI spec to 3.1.* spec"
|
||||
+ " for better support."
|
||||
)
|
||||
swagger_version = obj.get("swagger")
|
||||
openapi_version = obj.get("openapi")
|
||||
if isinstance(openapi_version, str):
|
||||
if openapi_version != "3.1.0":
|
||||
logger.warning(
|
||||
f"Attempting to load an OpenAPI {openapi_version}"
|
||||
f" spec. {warning_message}"
|
||||
)
|
||||
else:
|
||||
pass
|
||||
elif isinstance(swagger_version, str):
|
||||
logger.warning(
|
||||
f"Attempting to load a Swagger {swagger_version}"
|
||||
f" spec. {warning_message}"
|
||||
)
|
||||
else:
|
||||
raise ValueError(
|
||||
"Attempting to load an unsupported spec:"
|
||||
f"\n\n{obj}\n{warning_message}"
|
||||
)
|
||||
|
||||
@classmethod
|
||||
def parse_obj(cls, obj: dict) -> "OpenAPISpec":
|
||||
try:
|
||||
cls._alert_unsupported_spec(obj)
|
||||
return super().parse_obj(obj)
|
||||
except ValidationError as e:
|
||||
# We are handling possibly misconfigured specs and want to do a best-effort
|
||||
# job to get a reasonable interface out of it.
|
||||
new_obj = copy.deepcopy(obj)
|
||||
for error in e.errors():
|
||||
keys = error["loc"]
|
||||
item = new_obj
|
||||
for key in keys[:-1]:
|
||||
item = item[key]
|
||||
item.pop(keys[-1], None)
|
||||
return cls.parse_obj(new_obj)
|
||||
|
||||
@classmethod
|
||||
def from_spec_dict(cls, spec_dict: dict) -> "OpenAPISpec":
|
||||
"""Get an OpenAPI spec from a dict."""
|
||||
return cls.parse_obj(spec_dict)
|
||||
|
||||
@classmethod
|
||||
def from_text(cls, text: str) -> "OpenAPISpec":
|
||||
"""Get an OpenAPI spec from a text."""
|
||||
try:
|
||||
spec_dict = json.loads(text)
|
||||
except json.JSONDecodeError:
|
||||
spec_dict = yaml.safe_load(text)
|
||||
return cls.from_spec_dict(spec_dict)
|
||||
|
||||
@classmethod
|
||||
def from_file(cls, path: Union[str, Path]) -> "OpenAPISpec":
|
||||
"""Get an OpenAPI spec from a file path."""
|
||||
path_ = path if isinstance(path, Path) else Path(path)
|
||||
if not path_.exists():
|
||||
raise FileNotFoundError(f"{path} does not exist")
|
||||
with path_.open("r") as f:
|
||||
return cls.from_text(f.read())
|
||||
|
||||
@classmethod
|
||||
def from_url(cls, url: str) -> "OpenAPISpec":
|
||||
"""Get an OpenAPI spec from a URL."""
|
||||
response = requests.get(url)
|
||||
return cls.from_text(response.text)
|
||||
|
||||
@property
|
||||
def base_url(self) -> str:
|
||||
"""Get the base url."""
|
||||
return self.servers[0].url
|
||||
|
||||
def get_methods_for_path(self, path: str) -> List[str]:
|
||||
"""Return a list of valid methods for the specified path."""
|
||||
path_item = self._get_path_strict(path)
|
||||
results = []
|
||||
for method in HTTPVerb:
|
||||
operation = getattr(path_item, method.value, None)
|
||||
if isinstance(operation, Operation):
|
||||
results.append(method.value)
|
||||
return results
|
||||
|
||||
def get_operation(self, path: str, method: str) -> Operation:
|
||||
"""Get the operation object for a given path and HTTP method."""
|
||||
path_item = self._get_path_strict(path)
|
||||
operation_obj = getattr(path_item, method, None)
|
||||
if not isinstance(operation_obj, Operation):
|
||||
raise ValueError(f"No {method} method found for {path}")
|
||||
return operation_obj
|
||||
|
||||
def get_parameters_for_operation(self, operation: Operation) -> List[Parameter]:
|
||||
"""Get the components for a given operation."""
|
||||
parameters = []
|
||||
if operation.parameters:
|
||||
for parameter in operation.parameters:
|
||||
if isinstance(parameter, Reference):
|
||||
parameter = self._get_root_referenced_parameter(parameter)
|
||||
parameters.append(parameter)
|
||||
return parameters
|
||||
|
||||
def get_request_body_for_operation(
|
||||
self, operation: Operation
|
||||
) -> Optional[RequestBody]:
|
||||
"""Get the request body for a given operation."""
|
||||
request_body = operation.requestBody
|
||||
if isinstance(request_body, Reference):
|
||||
request_body = self._get_root_referenced_request_body(request_body)
|
||||
return request_body
|
||||
|
||||
@staticmethod
|
||||
def get_cleaned_operation_id(operation: Operation, path: str, method: str) -> str:
|
||||
"""Get a cleaned operation id from an operation id."""
|
||||
operation_id = operation.operationId
|
||||
if operation_id is None:
|
||||
# Replace all punctuation of any kind with underscore
|
||||
path = re.sub(r"[^a-zA-Z0-9]", "_", path.lstrip("/"))
|
||||
operation_id = f"{path}_{method}"
|
||||
return operation_id.replace("-", "_").replace(".", "_").replace("/", "_")
|
||||
|
||||
@@ -1,285 +0,0 @@
|
||||
"""Utility functions for parsing an OpenAPI spec."""
|
||||
import copy
|
||||
import json
|
||||
import logging
|
||||
import re
|
||||
from enum import Enum
|
||||
from pathlib import Path
|
||||
from typing import Dict, List, Optional, Union
|
||||
|
||||
import requests
|
||||
import yaml
|
||||
from openapi_schema_pydantic import (
|
||||
Components,
|
||||
OpenAPI,
|
||||
Operation,
|
||||
Parameter,
|
||||
PathItem,
|
||||
Paths,
|
||||
Reference,
|
||||
RequestBody,
|
||||
Schema,
|
||||
)
|
||||
from pydantic import ValidationError
|
||||
|
||||
logger = logging.getLogger(__name__)
|
||||
|
||||
|
||||
class HTTPVerb(str, Enum):
|
||||
"""HTTP verbs."""
|
||||
|
||||
GET = "get"
|
||||
PUT = "put"
|
||||
POST = "post"
|
||||
DELETE = "delete"
|
||||
OPTIONS = "options"
|
||||
HEAD = "head"
|
||||
PATCH = "patch"
|
||||
TRACE = "trace"
|
||||
|
||||
@classmethod
|
||||
def from_str(cls, verb: str) -> "HTTPVerb":
|
||||
"""Parse an HTTP verb."""
|
||||
try:
|
||||
return cls(verb)
|
||||
except ValueError:
|
||||
raise ValueError(f"Invalid HTTP verb. Valid values are {cls.__members__}")
|
||||
|
||||
|
||||
class OpenAPISpec(OpenAPI):
|
||||
"""OpenAPI Model that removes misformatted parts of the spec."""
|
||||
|
||||
@property
|
||||
def _paths_strict(self) -> Paths:
|
||||
if not self.paths:
|
||||
raise ValueError("No paths found in spec")
|
||||
return self.paths
|
||||
|
||||
def _get_path_strict(self, path: str) -> PathItem:
|
||||
path_item = self._paths_strict.get(path)
|
||||
if not path_item:
|
||||
raise ValueError(f"No path found for {path}")
|
||||
return path_item
|
||||
|
||||
@property
|
||||
def _components_strict(self) -> Components:
|
||||
"""Get components or err."""
|
||||
if self.components is None:
|
||||
raise ValueError("No components found in spec. ")
|
||||
return self.components
|
||||
|
||||
@property
|
||||
def _parameters_strict(self) -> Dict[str, Union[Parameter, Reference]]:
|
||||
"""Get parameters or err."""
|
||||
parameters = self._components_strict.parameters
|
||||
if parameters is None:
|
||||
raise ValueError("No parameters found in spec. ")
|
||||
return parameters
|
||||
|
||||
@property
|
||||
def _schemas_strict(self) -> Dict[str, Schema]:
|
||||
"""Get the dictionary of schemas or err."""
|
||||
schemas = self._components_strict.schemas
|
||||
if schemas is None:
|
||||
raise ValueError("No schemas found in spec. ")
|
||||
return schemas
|
||||
|
||||
@property
|
||||
def _request_bodies_strict(self) -> Dict[str, Union[RequestBody, Reference]]:
|
||||
"""Get the request body or err."""
|
||||
request_bodies = self._components_strict.requestBodies
|
||||
if request_bodies is None:
|
||||
raise ValueError("No request body found in spec. ")
|
||||
return request_bodies
|
||||
|
||||
def _get_referenced_parameter(self, ref: Reference) -> Union[Parameter, Reference]:
|
||||
"""Get a parameter (or nested reference) or err."""
|
||||
ref_name = ref.ref.split("/")[-1]
|
||||
parameters = self._parameters_strict
|
||||
if ref_name not in parameters:
|
||||
raise ValueError(f"No parameter found for {ref_name}")
|
||||
return parameters[ref_name]
|
||||
|
||||
def _get_root_referenced_parameter(self, ref: Reference) -> Parameter:
|
||||
"""Get the root reference or err."""
|
||||
parameter = self._get_referenced_parameter(ref)
|
||||
while isinstance(parameter, Reference):
|
||||
parameter = self._get_referenced_parameter(parameter)
|
||||
return parameter
|
||||
|
||||
def get_referenced_schema(self, ref: Reference) -> Schema:
|
||||
"""Get a schema (or nested reference) or err."""
|
||||
ref_name = ref.ref.split("/")[-1]
|
||||
schemas = self._schemas_strict
|
||||
if ref_name not in schemas:
|
||||
raise ValueError(f"No schema found for {ref_name}")
|
||||
return schemas[ref_name]
|
||||
|
||||
def get_schema(self, schema: Union[Reference, Schema]) -> Schema:
|
||||
if isinstance(schema, Reference):
|
||||
return self.get_referenced_schema(schema)
|
||||
return schema
|
||||
|
||||
def _get_root_referenced_schema(self, ref: Reference) -> Schema:
|
||||
"""Get the root reference or err."""
|
||||
schema = self.get_referenced_schema(ref)
|
||||
while isinstance(schema, Reference):
|
||||
schema = self.get_referenced_schema(schema)
|
||||
return schema
|
||||
|
||||
def _get_referenced_request_body(
|
||||
self, ref: Reference
|
||||
) -> Optional[Union[Reference, RequestBody]]:
|
||||
"""Get a request body (or nested reference) or err."""
|
||||
ref_name = ref.ref.split("/")[-1]
|
||||
request_bodies = self._request_bodies_strict
|
||||
if ref_name not in request_bodies:
|
||||
raise ValueError(f"No request body found for {ref_name}")
|
||||
return request_bodies[ref_name]
|
||||
|
||||
def _get_root_referenced_request_body(
|
||||
self, ref: Reference
|
||||
) -> Optional[RequestBody]:
|
||||
"""Get the root request Body or err."""
|
||||
request_body = self._get_referenced_request_body(ref)
|
||||
while isinstance(request_body, Reference):
|
||||
request_body = self._get_referenced_request_body(request_body)
|
||||
return request_body
|
||||
|
||||
@staticmethod
|
||||
def _alert_unsupported_spec(obj: dict) -> None:
|
||||
"""Alert if the spec is not supported."""
|
||||
warning_message = (
|
||||
" This may result in degraded performance."
|
||||
+ " Convert your OpenAPI spec to 3.1.* spec"
|
||||
+ " for better support."
|
||||
)
|
||||
swagger_version = obj.get("swagger")
|
||||
openapi_version = obj.get("openapi")
|
||||
if isinstance(openapi_version, str):
|
||||
if openapi_version != "3.1.0":
|
||||
logger.warning(
|
||||
f"Attempting to load an OpenAPI {openapi_version}"
|
||||
f" spec. {warning_message}"
|
||||
)
|
||||
else:
|
||||
pass
|
||||
elif isinstance(swagger_version, str):
|
||||
logger.warning(
|
||||
f"Attempting to load a Swagger {swagger_version}"
|
||||
f" spec. {warning_message}"
|
||||
)
|
||||
else:
|
||||
raise ValueError(
|
||||
"Attempting to load an unsupported spec:"
|
||||
f"\n\n{obj}\n{warning_message}"
|
||||
)
|
||||
|
||||
@classmethod
|
||||
def parse_obj(cls, obj: dict) -> "OpenAPISpec":
|
||||
try:
|
||||
cls._alert_unsupported_spec(obj)
|
||||
return super().parse_obj(obj)
|
||||
except ValidationError as e:
|
||||
# We are handling possibly misconfigured specs and want to do a best-effort
|
||||
# job to get a reasonable interface out of it.
|
||||
new_obj = copy.deepcopy(obj)
|
||||
for error in e.errors():
|
||||
keys = error["loc"]
|
||||
item = new_obj
|
||||
for key in keys[:-1]:
|
||||
item = item[key]
|
||||
item.pop(keys[-1], None)
|
||||
return cls.parse_obj(new_obj)
|
||||
|
||||
@classmethod
|
||||
def from_spec_dict(cls, spec_dict: dict) -> "OpenAPISpec":
|
||||
"""Get an OpenAPI spec from a dict."""
|
||||
return cls.parse_obj(spec_dict)
|
||||
|
||||
@classmethod
|
||||
def from_text(cls, text: str) -> "OpenAPISpec":
|
||||
"""Get an OpenAPI spec from a text."""
|
||||
try:
|
||||
spec_dict = json.loads(text)
|
||||
except json.JSONDecodeError:
|
||||
spec_dict = yaml.safe_load(text)
|
||||
return cls.from_spec_dict(spec_dict)
|
||||
|
||||
@classmethod
|
||||
def from_file(cls, path: Union[str, Path]) -> "OpenAPISpec":
|
||||
"""Get an OpenAPI spec from a file path."""
|
||||
path_ = path if isinstance(path, Path) else Path(path)
|
||||
if not path_.exists():
|
||||
raise FileNotFoundError(f"{path} does not exist")
|
||||
with path_.open("r") as f:
|
||||
return cls.from_text(f.read())
|
||||
|
||||
@classmethod
|
||||
def from_url(cls, url: str) -> "OpenAPISpec":
|
||||
"""Get an OpenAPI spec from a URL."""
|
||||
response = requests.get(url)
|
||||
return cls.from_text(response.text)
|
||||
|
||||
@property
|
||||
def base_url(self) -> str:
|
||||
"""Get the base url."""
|
||||
return self.servers[0].url
|
||||
|
||||
def get_methods_for_path(self, path: str) -> List[str]:
|
||||
"""Return a list of valid methods for the specified path."""
|
||||
path_item = self._get_path_strict(path)
|
||||
results = []
|
||||
for method in HTTPVerb:
|
||||
operation = getattr(path_item, method.value, None)
|
||||
if isinstance(operation, Operation):
|
||||
results.append(method.value)
|
||||
return results
|
||||
|
||||
def get_parameters_for_path(self, path: str) -> List[Parameter]:
|
||||
path_item = self._get_path_strict(path)
|
||||
parameters = []
|
||||
if not path_item.parameters:
|
||||
return []
|
||||
for parameter in path_item.parameters:
|
||||
if isinstance(parameter, Reference):
|
||||
parameter = self._get_root_referenced_parameter(parameter)
|
||||
parameters.append(parameter)
|
||||
return parameters
|
||||
|
||||
def get_operation(self, path: str, method: str) -> Operation:
|
||||
"""Get the operation object for a given path and HTTP method."""
|
||||
path_item = self._get_path_strict(path)
|
||||
operation_obj = getattr(path_item, method, None)
|
||||
if not isinstance(operation_obj, Operation):
|
||||
raise ValueError(f"No {method} method found for {path}")
|
||||
return operation_obj
|
||||
|
||||
def get_parameters_for_operation(self, operation: Operation) -> List[Parameter]:
|
||||
"""Get the components for a given operation."""
|
||||
parameters = []
|
||||
if operation.parameters:
|
||||
for parameter in operation.parameters:
|
||||
if isinstance(parameter, Reference):
|
||||
parameter = self._get_root_referenced_parameter(parameter)
|
||||
parameters.append(parameter)
|
||||
return parameters
|
||||
|
||||
def get_request_body_for_operation(
|
||||
self, operation: Operation
|
||||
) -> Optional[RequestBody]:
|
||||
"""Get the request body for a given operation."""
|
||||
request_body = operation.requestBody
|
||||
if isinstance(request_body, Reference):
|
||||
request_body = self._get_root_referenced_request_body(request_body)
|
||||
return request_body
|
||||
|
||||
@staticmethod
|
||||
def get_cleaned_operation_id(operation: Operation, path: str, method: str) -> str:
|
||||
"""Get a cleaned operation id from an operation id."""
|
||||
operation_id = operation.operationId
|
||||
if operation_id is None:
|
||||
# Replace all punctuation of any kind with underscore
|
||||
path = re.sub(r"[^a-zA-Z0-9]", "_", path.lstrip("/"))
|
||||
operation_id = f"{path}_{method}"
|
||||
return operation_id.replace("-", "_").replace(".", "_").replace("/", "_")
|
||||
10
poetry.lock
generated
10
poetry.lock
generated
@@ -4362,13 +4362,13 @@ tests = ["doctest", "pytest", "pytest-mock"]
|
||||
|
||||
[[package]]
|
||||
name = "langchainplus-sdk"
|
||||
version = "0.0.17"
|
||||
description = "Client library to connect to the LangSmith LLM Tracing and Evaluation Platform."
|
||||
version = "0.0.15"
|
||||
description = "Client library to connect to the LangChainPlus LLM Tracing and Evaluation Platform."
|
||||
optional = false
|
||||
python-versions = ">=3.8.1,<4.0"
|
||||
files = [
|
||||
{file = "langchainplus_sdk-0.0.17-py3-none-any.whl", hash = "sha256:899675fe850bb0829691ce7643d5c3b4425de1535b6f2d6ce1e5f5457ffb05bf"},
|
||||
{file = "langchainplus_sdk-0.0.17.tar.gz", hash = "sha256:6520c864a23dcadbe6fb7233a117347f6acc32725a97758e59354704c50de303"},
|
||||
{file = "langchainplus_sdk-0.0.15-py3-none-any.whl", hash = "sha256:e69bdbc8af6007ef2f774248d2483bbaf2d75712b1acc9ea50eda3b9f6dc567d"},
|
||||
{file = "langchainplus_sdk-0.0.15.tar.gz", hash = "sha256:ce40e9e3b6d42741f0a2aa89f83a12f2648f38690a9dd57e5fe3a56f2f232908"},
|
||||
]
|
||||
|
||||
[package.dependencies]
|
||||
@@ -11771,4 +11771,4 @@ text-helpers = ["chardet"]
|
||||
[metadata]
|
||||
lock-version = "2.0"
|
||||
python-versions = ">=3.8.1,<4.0"
|
||||
content-hash = "6e495e4f58127a5d2001385404b973896e275f5ca71a6ebe856cb114977189d1"
|
||||
content-hash = "09d46ad12369c6a16513558618553623cd520c2855bff3b8fe8248e1b18cbb94"
|
||||
|
||||
@@ -1,6 +1,6 @@
|
||||
[tool.poetry]
|
||||
name = "langchain"
|
||||
version = "0.0.211"
|
||||
version = "0.0.209"
|
||||
description = "Building applications with LLMs through composability"
|
||||
authors = []
|
||||
license = "MIT"
|
||||
@@ -106,7 +106,7 @@ pyspark = {version = "^3.4.0", optional = true}
|
||||
clarifai = {version = "9.1.0", optional = true}
|
||||
tigrisdb = {version = "^1.0.0b6", optional = true}
|
||||
nebula3-python = {version = "^3.4.0", optional = true}
|
||||
langchainplus-sdk = ">=0.0.17"
|
||||
langchainplus-sdk = ">=0.0.13"
|
||||
awadb = {version = "^0.3.3", optional = true}
|
||||
azure-search-documents = {version = "11.4.0a20230509004", source = "azure-sdk-dev", optional = true}
|
||||
openllm = {version = ">=0.1.6", optional = true}
|
||||
|
||||
60
scripts/clean_file.py
Normal file
60
scripts/clean_file.py
Normal file
@@ -0,0 +1,60 @@
|
||||
import argparse
|
||||
from typing import Optional, Sequence
|
||||
|
||||
from langchain.base_language import BaseLanguageModel
|
||||
from langchain.chains.llm import LLMChain
|
||||
from langchain.chat_models import ChatAnthropic, ChatOpenAI
|
||||
|
||||
|
||||
def parse_args(src: Optional[Sequence[str]] = None) -> argparse.Namespace:
|
||||
"""Parse arguments."""
|
||||
parser = argparse.ArgumentParser(
|
||||
description="Apply an LLM to a code file to proofread docstrings and edit grammar.",
|
||||
)
|
||||
parser.add_argument("file", type=str, help="File to proofread.")
|
||||
parser.add_argument(
|
||||
"--model",
|
||||
type=str,
|
||||
default="auto",
|
||||
help="Model to use.",
|
||||
choices=["anthropic", "openai", "auto"],
|
||||
)
|
||||
return parser.parse_args(src)
|
||||
|
||||
|
||||
def select_model(text: str, model: str) -> BaseLanguageModel:
|
||||
if model == "anthropic":
|
||||
return ChatAnthropic(model="claude-v1-100k", temperature=0)
|
||||
elif model == "openai":
|
||||
return ChatOpenAI(model="gpt-3.5-turbo-16k", temperature=0)
|
||||
elif model == "auto":
|
||||
import tiktoken
|
||||
|
||||
encoding = tiktoken.encoding_for_model("gpt-3.5-turbo-16k")
|
||||
num_tokens = len(encoding.encode(text))
|
||||
if num_tokens > 15800:
|
||||
return ChatAnthropic(model="claude-v1-100k", temperature=0)
|
||||
else:
|
||||
return ChatOpenAI(model="gpt-3.5-turbo-16k", temperature=0)
|
||||
else:
|
||||
raise ValueError(f"Invalid model {model}")
|
||||
|
||||
|
||||
def main(file: str, model: str) -> str:
|
||||
"""Run the llm."""
|
||||
with open(file, "r") as f:
|
||||
text = f.read()
|
||||
model_ = select_model(text, model)
|
||||
template = """Please review the following code and improve the docstrings to make our documentation clean. Update the docstrings and descriptions as needed in the format of scikit-learn. Provide examples if necessary.
|
||||
```
|
||||
{code}
|
||||
```"""
|
||||
chain = LLMChain.from_string(llm=model_, template=template)
|
||||
cleaned = chain(text, return_only_outputs=True)["text"]
|
||||
return cleaned.strip().strip("`")
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
args = parse_args()
|
||||
cleaned = main(args.file, args.model)
|
||||
print(cleaned)
|
||||
@@ -176,7 +176,7 @@ async def test_arun_on_dataset(monkeypatch: pytest.MonkeyPatch) -> None:
|
||||
{"result": f"Result for example {example.id}"} for _ in range(n_repetitions)
|
||||
]
|
||||
|
||||
def mock_create_project(*args: Any, **kwargs: Any) -> None:
|
||||
def mock_create_session(*args: Any, **kwargs: Any) -> None:
|
||||
pass
|
||||
|
||||
with mock.patch.object(
|
||||
@@ -186,7 +186,7 @@ async def test_arun_on_dataset(monkeypatch: pytest.MonkeyPatch) -> None:
|
||||
), mock.patch(
|
||||
"langchain.client.runner_utils._arun_llm_or_chain", new=mock_arun_chain
|
||||
), mock.patch.object(
|
||||
LangChainPlusClient, "create_project", new=mock_create_project
|
||||
LangChainPlusClient, "create_session", new=mock_create_session
|
||||
):
|
||||
client = LangChainPlusClient(api_url="http://localhost:1984", api_key="123")
|
||||
chain = mock.MagicMock()
|
||||
@@ -195,7 +195,7 @@ async def test_arun_on_dataset(monkeypatch: pytest.MonkeyPatch) -> None:
|
||||
dataset_name="test",
|
||||
llm_or_chain_factory=lambda: chain,
|
||||
concurrency_level=2,
|
||||
project_name="test_project",
|
||||
session_name="test_session",
|
||||
num_repetitions=num_repetitions,
|
||||
client=client,
|
||||
)
|
||||
|
||||
@@ -19,7 +19,6 @@ from langchain.tools.base import (
|
||||
StructuredTool,
|
||||
ToolException,
|
||||
)
|
||||
from tests.unit_tests.callbacks.fake_callback_handler import FakeCallbackHandler
|
||||
|
||||
|
||||
def test_unnamed_decorator() -> None:
|
||||
@@ -394,64 +393,6 @@ def test_empty_args_decorator() -> None:
|
||||
assert empty_tool_input.run({}) == "the empty result"
|
||||
|
||||
|
||||
def test_tool_from_function_with_run_manager() -> None:
|
||||
"""Test run of tool when using run_manager."""
|
||||
|
||||
def foo(bar: str, callbacks: Optional[CallbackManagerForToolRun] = None) -> str:
|
||||
"""Docstring
|
||||
Args:
|
||||
bar: str
|
||||
"""
|
||||
assert callbacks is not None
|
||||
return "foo" + bar
|
||||
|
||||
handler = FakeCallbackHandler()
|
||||
tool = Tool.from_function(foo, name="foo", description="Docstring")
|
||||
|
||||
assert tool.run(tool_input={"bar": "bar"}, run_manager=[handler]) == "foobar"
|
||||
assert tool.run("baz", run_manager=[handler]) == "foobaz"
|
||||
|
||||
|
||||
def test_structured_tool_from_function_with_run_manager() -> None:
|
||||
"""Test args and schema of structured tool when using callbacks."""
|
||||
|
||||
def foo(
|
||||
bar: int, baz: str, callbacks: Optional[CallbackManagerForToolRun] = None
|
||||
) -> str:
|
||||
"""Docstring
|
||||
Args:
|
||||
bar: int
|
||||
baz: str
|
||||
"""
|
||||
assert callbacks is not None
|
||||
return str(bar) + baz
|
||||
|
||||
handler = FakeCallbackHandler()
|
||||
structured_tool = StructuredTool.from_function(foo)
|
||||
|
||||
assert structured_tool.args == {
|
||||
"bar": {"title": "Bar", "type": "integer"},
|
||||
"baz": {"title": "Baz", "type": "string"},
|
||||
}
|
||||
|
||||
assert structured_tool.args_schema.schema() == {
|
||||
"properties": {
|
||||
"bar": {"title": "Bar", "type": "integer"},
|
||||
"baz": {"title": "Baz", "type": "string"},
|
||||
},
|
||||
"title": "fooSchemaSchema",
|
||||
"type": "object",
|
||||
"required": ["bar", "baz"],
|
||||
}
|
||||
|
||||
assert (
|
||||
structured_tool.run(
|
||||
tool_input={"bar": "10", "baz": "baz"}, run_manger=[handler]
|
||||
)
|
||||
== "10baz"
|
||||
)
|
||||
|
||||
|
||||
def test_named_tool_decorator() -> None:
|
||||
"""Test functionality when arguments are provided as input to decorator."""
|
||||
|
||||
|
||||
Reference in New Issue
Block a user