sunholo 0.118.10__py3-none-any.whl → 0.119.3__py3-none-any.whl

sunholo/discovery_engine/chunker_handler.py

@@ -65,8 +65,9 @@ def do_discovery_engine(message_data:str, metadata:dict, config:ConfigManager=No
             return None
         for corp in corpuses:
             try:
-                response = corp.import_documents(
-                    gcs_uri=message_data
+                response = corp.import_document_with_metadata(
+                    gcs_uri=message_data,
+                    metadata=metadata
                 )
                 log.info(f"Imported file to corpus: {response} with metadata: {metadata}")
             except Exception as err:
@@ -81,8 +82,9 @@ def do_discovery_engine(message_data:str, metadata:dict, config:ConfigManager=No
                         continue
                     if new_corp:
                         log.info(f"Found new DiscoveryEngine {config.vector_name=} - {new_corp=}")
-                        response = corp.import_documents(
-                            gcs_uri=message_data
+                        response = corp.import_document_with_metadata(
+                            gcs_uri=message_data,
+                            metadata=metadata
                         )
                     
                 continue

sunholo/discovery_engine/create_new.py

@@ -1,6 +1,7 @@
 from .discovery_engine_client import DiscoveryEngineClient
 from ..utils import ConfigManager
 from ..utils.gcp_project import get_gcp_project
+from ..custom_logging import log
 
 def create_new_discovery_engine(config:ConfigManager):
 
@@ -12,9 +13,13 @@ def create_new_discovery_engine(config:ConfigManager):
             chunk_size = chunker_config["chunk_size"]    
 
     gcp_config = config.vacConfig("gcp_config")
-    project_id = gcp_config.get("project_id") or get_gcp_project()
+    if not gcp_config:
+        log.info("Found no gcp_config in configuration so using get_gcp_project()")
+        project_id = get_gcp_project()
+    else:
+        project_id = gcp_config.get("project_id") or get_gcp_project()
     if not project_id:
-        raise ValueError("Could not find project_id in gcp_config")
+        raise ValueError("Could not find project_id in gcp_config or global")
     
     #location = gcp_config.get('location')
 

sunholo/discovery_engine/discovery_engine_client.py

@@ -1,6 +1,6 @@
 try:
     from google.api_core.client_options import ClientOptions
-    from google.cloud import discoveryengine_v1alpha as discoveryengine
+    from google.cloud import discoveryengine
     from google.api_core.retry import Retry, if_exception_type
     from google.api_core.exceptions import ResourceExhausted, AlreadyExists
     from google.cloud.discoveryengine_v1alpha import SearchResponse, Chunk
@@ -13,6 +13,9 @@ except ImportError:
 from ..custom_logging import log
 from typing import Optional, List
 import asyncio
+import json
+import uuid
+from ..utils.mime import guess_mime_type
 
 class DiscoveryEngineClient:
     """
@@ -431,7 +434,37 @@ class DiscoveryEngineClient:
 
         return operation.operation.name
 
+    def _import_document_request(self,
+        request
+    ) -> str:
+        """
+        Handles the common logic for making an ImportDocumentsRequest, including retrying.
 
+        Args:
+            request (discoveryengine.ImportDocumentsRequest): The prepared request object.
+
+        Returns:
+            str: The operation name.
+        """
+        @self.my_retry()
+        def import_documents_with_retry(doc_client, request):
+            return doc_client.import_documents(request=request)
+        
+        try:
+            operation = import_documents_with_retry(self.doc_client, request)
+        except ResourceExhausted as e:
+            log.error(f"DiscoveryEngine Operation failed after retries due to quota exceeded: {e}")
+            raise e
+        except AlreadyExists as e:
+            # Extract relevant info from the request to log
+            gcs_uri = request.gcs_source.input_uris if request.gcs_source else None
+            bigquery_table = request.bigquery_source.table_id if request.bigquery_source else None
+            log.warning(f"DiscoveryEngine - Already exists: {e} - {gcs_uri=} {bigquery_table=}")
+        except Exception as e:
+            log.error(f"An unexpected DiscoveryEngine error occurred: {e}")
+            raise e
+
+        return operation.operation.name
 
     def import_documents(self,
         gcs_uri: Optional[str] = None,
@@ -479,23 +512,115 @@ class DiscoveryEngineClient:
             )
 
         # Make the request
-        @self.my_retry()
-        def import_documents_with_retry(doc_client, request):
-            return doc_client.import_documents(request=request)
-        
-        try:
-            operation = import_documents_with_retry(self.doc_client, request)
-        except ResourceExhausted as e:
-            log.error(f"DiscoveryEngine Operation failed after retries due to quota exceeded: {e}")
+        return self._import_document_request(request)
 
-            raise e
-        except AlreadyExists as e:
-            log.warning(f"DiscoveryEngine - Already exists: {e} - {gcs_uri=} {bigquery_table=}")
 
-        except Exception as e:
-            log.error(f"An unexpected DiscoveryEngine error occurred: {e}")
+    def import_documents_with_metadata(self, gcs_uri: str, data_schema="content", branch="default_branch"):
+        """
+        Supply a JSONLD GCS location to import all the GS URIs within and their metadata
+        """
+        parent = self.doc_client.branch_path(
+            self.project_id, 
+            self.location, 
+            self.data_store_id, 
+            branch
+        )
 
-            raise e
+        # 1. Prepare your documents with metadata:
+        documents_with_metadata = []
+        with open(gcs_uri, 'r') as f:  # Assuming one JSON object per line in your GCS file
+            for line in f:
+                try:
+                    document_data = json.loads(line)  # Load the JSON from the line
+                    # Check if it has the required fields, if not create them
+                    if "id" not in document_data:
+                        document_data["id"] = str(uuid.uuid4())
+                    if "structData" not in document_data:
+                        document_data["structData"] = {}
+                    if "content" not in document_data:
+                        document_data["content"] = {}
+                    # Create the Document object with your metadata
+                    document = discoveryengine.Document(
+                        name = f"{parent}/documents/{document_data['id']}", # important!
+                        id=document_data["id"],
+                        struct_data=document_data.get("structData", {}),  # Your metadata here
+                        content = discoveryengine.Content(
+                            mime_type = document_data.get("content", {}).get("mimeType", "text/plain"),
+                            uri = document_data.get("content", {}).get("uri", ""),
+                        )
+                    )
+
+                    if "jsonData" in document_data:
+                        document.json_data = document_data["jsonData"]
+
+                    documents_with_metadata.append(document)
+
+                except json.JSONDecodeError as e:
+                    log.error(f"Error decoding JSON in line: {line.strip()}. Error: {e}")
+                    continue  # Skip to the next line if there's an error
+
+        # 2. Use InlineSource to import:
+        request = discoveryengine.ImportDocumentsRequest(
+            parent=parent,
+            inline_source=discoveryengine.ImportDocumentsRequest.InlineSource(
+                documents=documents_with_metadata,  # Pass the list of Document objects
+            ),
+            reconciliation_mode=discoveryengine.ImportDocumentsRequest.ReconciliationMode.INCREMENTAL,
+        )
 
-        return operation.operation.name
+        return self._import_document_request(request)
+
+    
+    def import_document_with_metadata(self, gcs_uri: str, metadata: dict, branch="default_branch"):
+        """
+        Imports a single document with metadata.
+
+        Args:
+            gcs_uri: The GCS URI of the document to import.
+            metadata: A dictionary containing the metadata for the document.
+            branch: The branch to import the document into.
 
+        Returns:
+            str: The operation name.
+        """
+        try:
+            # 1. Generate a unique document ID
+            document_id = str(uuid.uuid4())
+
+            # 2. Create a Document object
+            parent = self.doc_client.branch_path(
+                self.project_id, self.location, self.data_store_id, branch
+            )
+            document = discoveryengine.Document(
+                name=f"{parent}/documents/{document_id}",
+                id=document_id,
+                struct_data=metadata,
+                content=discoveryengine.Document.Content(
+                    uri=gcs_uri,
+                    mime_type=self.get_mime_type(gcs_uri)
+                )
+            )
+
+            # 3. Use InlineSource for import
+            request = discoveryengine.ImportDocumentsRequest(
+                parent=parent,
+                inline_source=discoveryengine.ImportDocumentsRequest.InlineSource(
+                    documents=[document],
+                ),
+                reconciliation_mode=discoveryengine.ImportDocumentsRequest.ReconciliationMode.INCREMENTAL
+            )
+
+            # 4. Make the import request (using the common method)
+            return self._import_document_request(request)
+
+        except Exception as e:
+            log.error(f"Error importing document with metadata: {e}")
+    
+    def get_mime_type(self, uri:str):
+        return guess_mime_type(uri)
+    
+    def search_with_filters(self, query, folder=None, date=None,
+                        num_previous_chunks=3, num_next_chunks=3, 
+                        page_size=10, parse_chunks_to_string=True, 
+                        serving_config="default_serving_config"):
+        pass

{sunholo-0.118.10.dist-info → sunholo-0.119.3.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: sunholo
-Version: 0.118.10
+Version: 0.119.3
 Summary: Large Language Model DevOps - a package to help deploy LLMs to the Cloud.
 Author-email: Holosun ApS <multivac@sunholo.com>
 License: Apache License, Version 2.0
@@ -47,7 +47,7 @@ Requires-Dist: google-cloud-service-control; extra == "all"
 Requires-Dist: google-cloud-logging; extra == "all"
 Requires-Dist: google-cloud-storage; extra == "all"
 Requires-Dist: google-cloud-pubsub; extra == "all"
-Requires-Dist: google-cloud-discoveryengine; extra == "all"
+Requires-Dist: google-cloud-discoveryengine>=0.13.4; extra == "all"
 Requires-Dist: google-cloud-texttospeech; extra == "all"
 Requires-Dist: google-generativeai>=0.7.1; extra == "all"
 Requires-Dist: google-genai; extra == "all"
@@ -133,7 +133,7 @@ Requires-Dist: google-cloud-service-control; extra == "gcp"
 Requires-Dist: google-cloud-storage; extra == "gcp"
 Requires-Dist: google-cloud-logging; extra == "gcp"
 Requires-Dist: google-cloud-pubsub; extra == "gcp"
-Requires-Dist: google-cloud-discoveryengine; extra == "gcp"
+Requires-Dist: google-cloud-discoveryengine>=0.13.4; extra == "gcp"
 Requires-Dist: google-cloud-texttospeech; extra == "gcp"
 Requires-Dist: google-genai; extra == "gcp"
 Requires-Dist: google-generativeai>=0.8.3; extra == "gcp"

{sunholo-0.118.10.dist-info → sunholo-0.119.3.dist-info}/RECORD

@@ -72,9 +72,9 @@ sunholo/database/sql/sb/delete_source_row.sql,sha256=r6fEuUKdbiLHCDGKSbKINDCpJjs
 sunholo/database/sql/sb/return_sources.sql,sha256=89KAnxfK8n_qGK9jy1OQT8f9n4uYUtYL5cCxbC2mj_c,255
 sunholo/database/sql/sb/setup.sql,sha256=CvoFvZQev2uWjmFa3aj3m3iuPFzAAJZ0S7Qi3L3-zZI,89
 sunholo/discovery_engine/__init__.py,sha256=hLgqRDJ22Aov9o2QjAEfsVgnL3kMdM-g5p8RJ9OyKdQ,130
-sunholo/discovery_engine/chunker_handler.py,sha256=5tw5_jalNQosf7uFyCmsYA__VdNpWC1PPVVa420CzWU,5479
-sunholo/discovery_engine/create_new.py,sha256=jWg5LW-QpFE8zq50ShaQJB3Wu8loiWB0P4lRWaCHpss,1023
-sunholo/discovery_engine/discovery_engine_client.py,sha256=jfIayVUOPM4svGF1S5Kk60rIG-xSo_e3zOHtBRg0nZA,22002
+sunholo/discovery_engine/chunker_handler.py,sha256=E3z-rVUuhjDebJY6nderr9QBYe8CrjwKskwIkOa_e68,5591
+sunholo/discovery_engine/create_new.py,sha256=WUi4_xh_dFaGX3xA9jkNKZhaR6LCELjMPeRb0hyj4FU,1226
+sunholo/discovery_engine/discovery_engine_client.py,sha256=C9fz341ZFMPtVSvqw2DbAgosJ5r5-YjfigRK-uFsldY,27407
 sunholo/discovery_engine/get_ai_search_chunks.py,sha256=hsFGOQugSeTMPEaQ16XTs_D45F8NABBm2IsAEdTk7kQ,4316
 sunholo/embedder/__init__.py,sha256=sI4N_CqgEVcrMDxXgxKp1FsfsB4FpjoXgPGkl4N_u4I,44
 sunholo/embedder/embed_chunk.py,sha256=Vvvj3-H4pSb1a2sLik3-X3X459j2jrUq1dBNAsOcQLo,7156
@@ -164,9 +164,9 @@ sunholo/vertex/init.py,sha256=1OQwcPBKZYBTDPdyU7IM4X4OmiXLdsNV30C-fee2scQ,2875
 sunholo/vertex/memory_tools.py,sha256=tBZxqVZ4InTmdBvLlOYwoSEWu4-kGquc-gxDwZCC4FA,7667
 sunholo/vertex/safety.py,sha256=S9PgQT1O_BQAkcqauWncRJaydiP8Q_Jzmu9gxYfy1VA,2482
 sunholo/vertex/type_dict_to_json.py,sha256=uTzL4o9tJRao4u-gJOFcACgWGkBOtqACmb6ihvCErL8,4694
-sunholo-0.118.10.dist-info/LICENSE.txt,sha256=SdE3QjnD3GEmqqg9EX3TM9f7WmtOzqS1KJve8rhbYmU,11345
-sunholo-0.118.10.dist-info/METADATA,sha256=prYJcbvo8lNaYqKHCeNC-I2_9ObUb8hr_AZHt6YMRRY,9639
-sunholo-0.118.10.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
-sunholo-0.118.10.dist-info/entry_points.txt,sha256=bZuN5AIHingMPt4Ro1b_T-FnQvZ3teBes-3OyO0asl4,49
-sunholo-0.118.10.dist-info/top_level.txt,sha256=wt5tadn5--5JrZsjJz2LceoUvcrIvxjHJe-RxuudxAk,8
-sunholo-0.118.10.dist-info/RECORD,,
+sunholo-0.119.3.dist-info/LICENSE.txt,sha256=SdE3QjnD3GEmqqg9EX3TM9f7WmtOzqS1KJve8rhbYmU,11345
+sunholo-0.119.3.dist-info/METADATA,sha256=OmzF8MtMCAYTxGe_VmAh8Zv9eMFwfwQ5ZC85d0FtZHg,9654
+sunholo-0.119.3.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
+sunholo-0.119.3.dist-info/entry_points.txt,sha256=bZuN5AIHingMPt4Ro1b_T-FnQvZ3teBes-3OyO0asl4,49
+sunholo-0.119.3.dist-info/top_level.txt,sha256=wt5tadn5--5JrZsjJz2LceoUvcrIvxjHJe-RxuudxAk,8
+sunholo-0.119.3.dist-info/RECORD,,