community[patch]: adding linearization config to AmazonTextractPDFLoader (#17489)

- **Description:** Adding an optional parameter `linearization_config`
to the `AmazonTextractPDFLoader` so the caller can define how the output
will be linearized, instead of forcing a predefined set of linearization
configs. It will still have a default configuration as this will be an
optional parameter.
- **Issue:** #17457
- **Dependencies:** The same ones that already exist for
`AmazonTextractPDFLoader`
- **Twitter handle:** [@lvieirajr19](https://twitter.com/lvieirajr19)

---------

Co-authored-by: Bagatur <baskaryan@gmail.com>

docs/docs/integrations/document_loaders/amazon_textract.ipynb

@@ -206,6 +206,42 @@
     "len(documents)"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "id": "a56ba97505c8d140",
+   "metadata": {
+    "collapsed": false,
+    "jupyter": {
+     "outputs_hidden": false
+    }
+   },
+   "source": [
+    "## Sample 4\n",
+    "\n",
+    "You have the option to pass an additional parameter called `linearization_config` to the AmazonTextractPDFLoader which will determine how the the text output will be linearized by the parser after Textract runs."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "1efbc4b6-f3cb-45c5-bbe8-16e7df060b92",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain_community.document_loaders import AmazonTextractPDFLoader\n",
+    "from textractor.data.text_linearization_config import TextLinearizationConfig\n",
+    "\n",
+    "loader = AmazonTextractPDFLoader(\n",
+    "    \"s3://amazon-textract-public-content/langchain/layout-parser-paper.pdf\",\n",
+    "    linearization_config=TextLinearizationConfig(\n",
+    "        hide_header_layout=True,\n",
+    "        hide_footer_layout=True,\n",
+    "        hide_figure_layout=True,\n",
+    "    ),\n",
+    ")\n",
+    "documents = loader.load()"
+   ]
+  },
   {
    "cell_type": "markdown",
    "id": "b3e41b4d-b159-4274-89be-80d8159134ef",
@@ -276,11 +312,14 @@
    ]
   },
   {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "1a09d18b-ab7b-468e-ae66-f92abf666b9b",
-   "metadata": {},
-   "outputs": [],
+   "cell_type": "markdown",
+   "id": "bd97f1c90aff6a83",
+   "metadata": {
+    "collapsed": false,
+    "jupyter": {
+     "outputs_hidden": false
+    }
+   },
    "source": []
   }
  ],
@@ -876,7 +915,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.10.12"
+   "version": "3.10.13"
   }
  },
  "nbformat": 4,

libs/community/langchain_community/document_loaders/parsers/pdf.py

@@ -26,6 +26,7 @@ if TYPE_CHECKING:
     import pdfplumber.page
     import pypdf._page
     import pypdfium2._helpers.page
+    from textractor.data.text_linearization_config import TextLinearizationConfig
 
 
 _PDF_FILTER_WITH_LOSS = ["DCTDecode", "DCT", "JPXDecode"]
@@ -454,6 +455,8 @@ class AmazonTextractPDFParser(BaseBlobParser):
         self,
         textract_features: Optional[Sequence[int]] = None,
         client: Optional[Any] = None,
+        *,
+        linearization_config: Optional["TextLinearizationConfig"] = None,
     ) -> None:
         """Initializes the parser.
 
@@ -462,6 +465,9 @@ class AmazonTextractPDFParser(BaseBlobParser):
                                should be passed as an int that conforms to the enum
                                `Textract_Features`, see `amazon-textract-caller` pkg
             client: boto3 textract client
+            linearization_config: Config to be used for linearization of the output
+                                  should be an instance of TextLinearizationConfig from
+                                  the `textractor` pkg
         """
 
         try:
@@ -477,6 +483,16 @@ class AmazonTextractPDFParser(BaseBlobParser):
                 ]
             else:
                 self.textract_features = []
+
+            if linearization_config is not None:
+                self.linearization_config = linearization_config
+            else:
+                self.linearization_config = self.textractor.TextLinearizationConfig(
+                    hide_figure_layout=True,
+                    title_prefix="# ",
+                    section_header_prefix="## ",
+                    list_element_prefix="*",
+                )
         except ImportError:
             raise ImportError(
                 "Could not import amazon-textract-caller or "
@@ -527,15 +543,9 @@ class AmazonTextractPDFParser(BaseBlobParser):
 
         document = self.textractor.Document.open(textract_response_json)
 
-        linearizer_config = self.textractor.TextLinearizationConfig(
-            hide_figure_layout=True,
-            title_prefix="# ",
-            section_header_prefix="## ",
-            list_element_prefix="*",
-        )
         for idx, page in enumerate(document.pages):
             yield Document(
-                page_content=page.get_text(config=linearizer_config),
+                page_content=page.get_text(config=self.linearization_config),
                 metadata={"source": blob.source, "page": idx + 1},
             )
 

libs/community/langchain_community/document_loaders/pdf.py

@@ -6,7 +6,17 @@ import time
 from abc import ABC
 from io import StringIO
 from pathlib import Path
-from typing import Any, Dict, Iterator, List, Mapping, Optional, Sequence, Union
+from typing import (
+    TYPE_CHECKING,
+    Any,
+    Dict,
+    Iterator,
+    List,
+    Mapping,
+    Optional,
+    Sequence,
+    Union,
+)
 from urllib.parse import urlparse
 
 import requests
@@ -26,6 +36,9 @@ from langchain_community.document_loaders.parsers.pdf import (
 )
 from langchain_community.document_loaders.unstructured import UnstructuredFileLoader
 
+if TYPE_CHECKING:
+    from textractor.data.text_linearization_config import TextLinearizationConfig
+
 logger = logging.getLogger(__file__)
 
 
@@ -586,6 +599,8 @@ class AmazonTextractPDFLoader(BasePDFLoader):
         region_name: Optional[str] = None,
         endpoint_url: Optional[str] = None,
         headers: Optional[Dict] = None,
+        *,
+        linearization_config: Optional["TextLinearizationConfig"] = None,
     ) -> None:
         """Initialize the loader.
 
@@ -598,7 +613,9 @@ class AmazonTextractPDFLoader(BasePDFLoader):
             credentials_profile_name: AWS profile name, if not default (Optional)
             region_name: AWS region, eg us-east-1 (Optional)
             endpoint_url: endpoint url for the textract service (Optional)
-
+            linearization_config: Config to be used for linearization of the output
+                                  should be an instance of TextLinearizationConfig from
+                                  the `textractor` pkg
         """
         super().__init__(file_path, headers=headers)
 
@@ -643,7 +660,11 @@ class AmazonTextractPDFLoader(BasePDFLoader):
                     "Please check that credentials in the specified "
                     "profile name are valid."
                 ) from e
-        self.parser = AmazonTextractPDFParser(textract_features=features, client=client)
+        self.parser = AmazonTextractPDFParser(
+            textract_features=features,
+            client=client,
+            linearization_config=linearization_config,
+        )
 
     def load(self) -> List[Document]:
         """Load given path as pages."""