sunholo 0.71.0__tar.gz → 0.71.3__tar.gz

{sunholo-0.71.0 → sunholo-0.71.3}/PKG-INFO

@@ -1,9 +1,9 @@
 Metadata-Version: 2.1
 Name: sunholo
-Version: 0.71.0
+Version: 0.71.3
 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/v0.71.0.tar.gz
+Download-URL: https://github.com/sunholo-data/sunholo-py/archive/refs/tags/v0.71.3.tar.gz
 Author: Holosun ApS
 Author-email: multivac@sunholo.com
 License: Apache License, Version 2.0
@@ -60,10 +60,13 @@ Requires-Dist: playwright; extra == "all"
 Requires-Dist: psycopg2-binary; extra == "all"
 Requires-Dist: pypdf; extra == "all"
 Requires-Dist: python-socketio; extra == "all"
+Requires-Dist: pytesseract; extra == "all"
 Requires-Dist: rich; extra == "all"
 Requires-Dist: supabase; extra == "all"
+Requires-Dist: tabulate; extra == "all"
 Requires-Dist: tantivy; extra == "all"
 Requires-Dist: tiktoken; extra == "all"
+Requires-Dist: unstructured[local-inference]; extra == "all"
 Provides-Extra: cli
 Requires-Dist: jsonschema>=4.21.1; extra == "cli"
 Requires-Dist: rich; extra == "cli"
@@ -80,7 +83,9 @@ Provides-Extra: pipeline
 Requires-Dist: GitPython; extra == "pipeline"
 Requires-Dist: lark; extra == "pipeline"
 Requires-Dist: pypdf; extra == "pipeline"
+Requires-Dist: pytesseract; extra == "pipeline"
 Requires-Dist: tabulate; extra == "pipeline"
+Requires-Dist: unstructured[local-inference]; extra == "pipeline"
 Provides-Extra: gcp
 Requires-Dist: google-auth-httplib2; extra == "gcp"
 Requires-Dist: google-auth-oauthlib; extra == "gcp"

{sunholo-0.71.0 → sunholo-0.71.3}/setup.py

@@ -1,7 +1,7 @@
 from setuptools import setup, find_packages
 
 # Define your base version
-version = '0.71.0'
+version = '0.71.3'
 
 setup(
     name='sunholo',
@@ -75,10 +75,13 @@ setup(
             "psycopg2-binary",
             "pypdf",
             "python-socketio",
+            "pytesseract",
             "rich",
             "supabase",
+            "tabulate",
             "tantivy",
             "tiktoken",
+            "unstructured[local-inference]",
 
         ],
         'cli': [
@@ -99,7 +102,9 @@ setup(
             "GitPython",
             "lark",
             "pypdf",
+            "pytesseract",
             "tabulate",
+            "unstructured[local-inference]",
         ],
         'gcp': [
             "google-auth-httplib2",

{sunholo-0.71.0 → sunholo-0.71.3}/sunholo/components/vectorstore.py

@@ -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}')   
+ 

{sunholo-0.71.0 → sunholo-0.71.3}/sunholo/llamaindex/import_files.py

@@ -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")

{sunholo-0.71.0 → sunholo-0.71.3}/sunholo/vertex/memory_tools.py

@@ -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-0.71.0 → sunholo-0.71.3}/sunholo.egg-info/PKG-INFO

@@ -1,9 +1,9 @@
 Metadata-Version: 2.1
 Name: sunholo
-Version: 0.71.0
+Version: 0.71.3
 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/v0.71.0.tar.gz
+Download-URL: https://github.com/sunholo-data/sunholo-py/archive/refs/tags/v0.71.3.tar.gz
 Author: Holosun ApS
 Author-email: multivac@sunholo.com
 License: Apache License, Version 2.0
@@ -60,10 +60,13 @@ Requires-Dist: playwright; extra == "all"
 Requires-Dist: psycopg2-binary; extra == "all"
 Requires-Dist: pypdf; extra == "all"
 Requires-Dist: python-socketio; extra == "all"
+Requires-Dist: pytesseract; extra == "all"
 Requires-Dist: rich; extra == "all"
 Requires-Dist: supabase; extra == "all"
+Requires-Dist: tabulate; extra == "all"
 Requires-Dist: tantivy; extra == "all"
 Requires-Dist: tiktoken; extra == "all"
+Requires-Dist: unstructured[local-inference]; extra == "all"
 Provides-Extra: cli
 Requires-Dist: jsonschema>=4.21.1; extra == "cli"
 Requires-Dist: rich; extra == "cli"
@@ -80,7 +83,9 @@ Provides-Extra: pipeline
 Requires-Dist: GitPython; extra == "pipeline"
 Requires-Dist: lark; extra == "pipeline"
 Requires-Dist: pypdf; extra == "pipeline"
+Requires-Dist: pytesseract; extra == "pipeline"
 Requires-Dist: tabulate; extra == "pipeline"
+Requires-Dist: unstructured[local-inference]; extra == "pipeline"
 Provides-Extra: gcp
 Requires-Dist: google-auth-httplib2; extra == "gcp"
 Requires-Dist: google-auth-oauthlib; extra == "gcp"

{sunholo-0.71.0 → sunholo-0.71.3}/sunholo.egg-info/requires.txt

@@ -41,10 +41,13 @@ playwright
 psycopg2-binary
 pypdf
 python-socketio
+pytesseract
 rich
 supabase
+tabulate
 tantivy
 tiktoken
+unstructured[local-inference]
 
 [anthropic]
 langchain-anthropic>=0.1.13
@@ -98,7 +101,9 @@ tiktoken
 GitPython
 lark
 pypdf
+pytesseract
 tabulate
+unstructured[local-inference]
 
 [tools]
 playwright