PyPI - sunholo - Versions diffs - 0.71.2__tar.gz → 0.71.4__tar.gz - Mend

@@ -1,9 +1,9 @@
 Metadata-Version: 2.1
 Name: sunholo
-Version: 0.71.2
+Version: 0.71.4
 Summary: Large Language Model DevOps - a package to help deploy LLMs to the Cloud.
 Home-page: https://github.com/sunholo-data/sunholo-py
-Download-URL: https://github.com/sunholo-data/sunholo-py/archive/refs/tags/v0.71.2.tar.gz
+Download-URL: https://github.com/sunholo-data/sunholo-py/archive/refs/tags/v0.71.4.tar.gz
 Author: Holosun ApS
 Author-email: multivac@sunholo.com
 License: Apache License, Version 2.0

@@ -1,7 +1,7 @@
 from setuptools import setup, find_packages
 # Define your base version
-version = '0.71.2'
+version = '0.71.4'
 setup(
     name='sunholo',

@@ -136,6 +136,4 @@ def pick_vectorstore(vs_str: str, vector_name: str, embeddings, read_only=None):
         log.info(f"Chose LanceDB for {vector_name} using {LANCEDB_BUCKET}")
         return vectorstore
-    else:
-        log.warning(f'No llm implemented for {vs_str}')

@@ -101,6 +101,11 @@ class DiscoveryEngineClient:
             str: The name of the long-running operation for data store creation.
         """
+        if chunk_size > 500:
+            chunk_size = 500
+        elif chunk_size < 100:
+            chunk_size = 100
         # https://cloud.google.com/python/docs/reference/discoveryengine/latest/google.cloud.discoveryengine_v1alpha.types.DocumentProcessingConfig
         doc_config = discoveryengine.DocumentProcessingConfig(
             chunking_config=discoveryengine.DocumentProcessingConfig.ChunkingConfig(

@@ -69,11 +69,16 @@ def do_llamaindex(message_data, metadata, vector_name):
                 rag_id = value.get('rag_id')
                 project_id = gcp_config.get('project_id')
                 location = gcp_config.get('location')
-                corpus = fetch_corpus(
-                    project_id=project_id or global_project_id,
-                    location=location or global_location,
-                    rag_id=rag_id or global_rag_id
-                )
+                try:
+                    corpus = fetch_corpus(
+                        project_id=project_id or global_project_id,
+                        location=location or global_location,
+                        rag_id=rag_id or global_rag_id
+                    )
+                except Exception as err:
+                    log.warning(f"Failed to fetch LlamaIndex corpus: {err=}")
+                    continue
                 corpuses.append(corpus)
     if not corpuses:
         log.error("Could not find a RAG corpus to import data to")

@@ -62,8 +62,8 @@ def get_vertex_memories(vector_name):
                 if rag_id is None:
                     raise ValueError("Must specify rag_id if using vectorstore: llamaindex")
-                project_id = gcp_config.get('project_id')
-                location = gcp_config.get('location')
+                project_id = value.get('project_id') or gcp_config.get('project_id')
+                location = value.get('location') or gcp_config.get('location')
                 try:
                     corpus = fetch_corpus(
@@ -87,7 +87,8 @@ def get_vertex_memories(vector_name):
             elif vectorstore == "discovery_engine" or vectorstore == "vertex_ai_search":
                 try:
-                    de = DiscoveryEngineClient(vector_name, project_id=get_gcp_project())
+                    project_id = value.get('project_id') or get_gcp_project()
+                    de = DiscoveryEngineClient(vector_name, project_id=project_id)
                     log.info(f"Found vectorstore {vectorstore}")
                     data_store_path = f"{de.data_store_path()}/dataStores/{vector_name}"

sunholo 0.71.2__tar.gz → 0.71.4__tar.gz