feat(API): Ingest plain text (#1417)

* Add ingest/text route to ingest plain text

* Add new ingest text test and adapt ingest/file ones

* Include new API in docs

* Remove duplicated logic

private_gpt/server/ingest/ingest_router.py

@@ -1,7 +1,7 @@
 from typing import Literal
 
 from fastapi import APIRouter, Depends, HTTPException, Request, UploadFile
-from pydantic import BaseModel
+from pydantic import BaseModel, Field
 
 from private_gpt.server.ingest.ingest_service import IngestService
 from private_gpt.server.ingest.model import IngestedDoc
@@ -10,14 +10,35 @@ from private_gpt.server.utils.auth import authenticated
 ingest_router = APIRouter(prefix="/v1", dependencies=[Depends(authenticated)])
 
 
+class IngestTextBody(BaseModel):
+    file_name: str = Field(examples=["Avatar: The Last Airbender"])
+    text: str = Field(
+        examples=[
+            "Avatar is set in an Asian and Arctic-inspired world in which some "
+            "people can telekinetically manipulate one of the four elements—water, "
+            "earth, fire or air—through practices known as 'bending', inspired by "
+            "Chinese martial arts."
+        ]
+    )
+
+
 class IngestResponse(BaseModel):
     object: Literal["list"]
     model: Literal["private-gpt"]
     data: list[IngestedDoc]
 
 
-@ingest_router.post("/ingest", tags=["Ingestion"])
+@ingest_router.post("/ingest", tags=["Ingestion"], deprecated=True)
 def ingest(request: Request, file: UploadFile) -> IngestResponse:
+    """Ingests and processes a file.
+
+    Deprecated. Use ingest/file instead.
+    """
+    return ingest_file(request, file)
+
+
+@ingest_router.post("/ingest/file", tags=["Ingestion"])
+def ingest_file(request: Request, file: UploadFile) -> IngestResponse:
     """Ingests and processes a file, storing its chunks to be used as context.
 
     The context obtained from files is later used in
@@ -40,6 +61,26 @@ def ingest(request: Request, file: UploadFile) -> IngestResponse:
     return IngestResponse(object="list", model="private-gpt", data=ingested_documents)
 
 
+@ingest_router.post("/ingest/text", tags=["Ingestion"])
+def ingest_text(request: Request, body: IngestTextBody) -> IngestResponse:
+    """Ingests and processes a text, storing its chunks to be used as context.
+
+    The context obtained from files is later used in
+    `/chat/completions`, `/completions`, and `/chunks` APIs.
+
+    A Document will be generated with the given text. The Document
+    ID is returned in the response, together with the
+    extracted Metadata (which is later used to improve context retrieval). That ID
+    can be used to filter the context used to create responses in
+    `/chat/completions`, `/completions`, and `/chunks` APIs.
+    """
+    service = request.state.injector.get(IngestService)
+    if len(body.file_name) == 0:
+        raise HTTPException(400, "No file name provided")
+    ingested_documents = service.ingest_text(body.file_name, body.text)
+    return IngestResponse(object="list", model="private-gpt", data=ingested_documents)
+
+
 @ingest_router.get("/ingest/list", tags=["Ingestion"])
 def list_ingested(request: Request) -> IngestResponse:
     """Lists already ingested Documents including their Document ID and metadata.

private_gpt/server/ingest/ingest_service.py

@@ -1,7 +1,7 @@
 import logging
 import tempfile
 from pathlib import Path
-from typing import BinaryIO
+from typing import AnyStr, BinaryIO
 
 from injector import inject, singleton
 from llama_index import (
@@ -53,16 +53,7 @@ class IngestService:
             self.storage_context, self.ingest_service_context, settings=settings()
         )
 
-    def ingest(self, file_name: str, file_data: Path) -> list[IngestedDoc]:
-        logger.info("Ingesting file_name=%s", file_name)
-        documents = self.ingest_component.ingest(file_name, file_data)
-        return [IngestedDoc.from_document(document) for document in documents]
-
-    def ingest_bin_data(
-        self, file_name: str, raw_file_data: BinaryIO
-    ) -> list[IngestedDoc]:
-        logger.debug("Ingesting binary data with file_name=%s", file_name)
-        file_data = raw_file_data.read()
+    def _ingest_data(self, file_name: str, file_data: AnyStr) -> list[IngestedDoc]:
         logger.debug("Got file data of size=%s to ingest", len(file_data))
         # llama-index mainly supports reading from files, so
         # we have to create a tmp file to read for it to work
@@ -74,11 +65,27 @@ class IngestService:
                     path_to_tmp.write_bytes(file_data)
                 else:
                     path_to_tmp.write_text(str(file_data))
-                return self.ingest(file_name, path_to_tmp)
+                return self.ingest_file(file_name, path_to_tmp)
             finally:
                 tmp.close()
                 path_to_tmp.unlink()
 
+    def ingest_file(self, file_name: str, file_data: Path) -> list[IngestedDoc]:
+        logger.info("Ingesting file_name=%s", file_name)
+        documents = self.ingest_component.ingest(file_name, file_data)
+        return [IngestedDoc.from_document(document) for document in documents]
+
+    def ingest_text(self, file_name: str, text: str) -> list[IngestedDoc]:
+        logger.debug("Ingesting text data with file_name=%s", file_name)
+        return self._ingest_data(file_name, text)
+
+    def ingest_bin_data(
+        self, file_name: str, raw_file_data: BinaryIO
+    ) -> list[IngestedDoc]:
+        logger.debug("Ingesting binary data with file_name=%s", file_name)
+        file_data = raw_file_data.read()
+        return self._ingest_data(file_name, file_data)
+
     def bulk_ingest(self, files: list[tuple[str, Path]]) -> list[IngestedDoc]:
         logger.info("Ingesting file_names=%s", [f[0] for f in files])
         documents = self.ingest_component.bulk_ingest(files)