sunholo 0.69.9__tar.gz → 0.69.12__tar.gz

{sunholo-0.69.9 → sunholo-0.69.12}/PKG-INFO

@@ -1,9 +1,9 @@
 Metadata-Version: 2.1
 Name: sunholo
-Version: 0.69.9
+Version: 0.69.12
 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/v0.69.9.tar.gz
+Download-URL: https://github.com/sunholo-data/sunholo-py/archive/refs/tags/v0.69.12.tar.gz
 Author: Holosun ApS
 Author-email: multivac@sunholo.com
 License: Apache License, Version 2.0
@@ -61,6 +61,7 @@ Requires-Dist: pypdf; extra == "all"
 Requires-Dist: python-socketio; extra == "all"
 Requires-Dist: rich; extra == "all"
 Requires-Dist: supabase; extra == "all"
+Requires-Dist: tantivy; extra == "all"
 Requires-Dist: tiktoken; extra == "all"
 Provides-Extra: cli
 Requires-Dist: jsonschema>=4.21.1; extra == "cli"
@@ -73,6 +74,7 @@ Requires-Dist: pg8000; extra == "database"
 Requires-Dist: pgvector; extra == "database"
 Requires-Dist: psycopg2-binary; extra == "database"
 Requires-Dist: lancedb; extra == "database"
+Requires-Dist: tantivy; extra == "database"
 Provides-Extra: pipeline
 Requires-Dist: GitPython; extra == "pipeline"
 Requires-Dist: lark; extra == "pipeline"

{sunholo-0.69.9 → sunholo-0.69.12}/setup.py

@@ -1,7 +1,7 @@
 from setuptools import setup, find_packages
 
 # Define your base version
-version = '0.69.9'
+version = '0.69.12'
 
 setup(
     name='sunholo',
@@ -76,6 +76,7 @@ setup(
             "python-socketio",
             "rich",
             "supabase",
+            "tantivy",
             "tiktoken",
 
         ],
@@ -91,6 +92,7 @@ setup(
             "pgvector",
             "psycopg2-binary",
             "lancedb",
+            "tantivy"
         ],
         'pipeline': [
             "GitPython",

{sunholo-0.69.9 → sunholo-0.69.12}/sunholo/discovery_engine/chunker_handler.py

@@ -51,6 +51,8 @@ def do_discovery_engine(message_data, metadata, vector_name):
 
     if message_data.startswith("gs://"):
         log.info(f"DiscoveryEngineClient.import_files for {message_data}")
+        if "/pdf_parts/" in message_data:
+            return None
         for corp in corpuses:
             try:
                 response = corp.import_documents(

{sunholo-0.69.9 → sunholo-0.69.12}/sunholo/discovery_engine/discovery_engine_client.py

@@ -155,6 +155,7 @@ class DiscoveryEngineClient:
         num_previous_chunks: int = 3,
         num_next_chunks: int = 3,
         page_size: int = 10,
+        parse_chunks_to_string: bool = True,
         doc_or_chunks: str = "CHUNKS",  # or DOCUMENTS
         serving_config: str = "default_serving_config",
     ):
@@ -166,6 +167,7 @@ class DiscoveryEngineClient:
             num_previous_chunks (int, optional): Number of previous chunks to return for context (default is 3).
             num_next_chunks (int, optional): Number of next chunks to return for context (default is 3).
             page_size (int, optional): The maximum number of results to return per page (default is 10).
+            parse_chunks_to_string: If True will put chunks in one big string, False will return object
 
         Returns:
             discoveryengine.SearchResponse: The search response object containing the search results.
@@ -199,10 +201,69 @@ class DiscoveryEngineClient:
             ),
         )
 
+        log.info(f"Discovery engine request: {search_request=}")
         search_response = self.search_client.search(search_request)
+        
+
+        if parse_chunks_to_string:
+
+            big_string = self.process_chunks(search_response)
+            log.info(f"Discovery engine chunks string sample: {big_string[:100]}")
 
+            return big_string
+        
+        log.info("Discovery engine response object")
         return search_response
 
+    def process_chunks(self, response):
+        all_chunks = []
+
+        if 'results' not in response:
+            raise ValueError(f'No results found in response: {response=}')
+        
+        for result in response['results']:
+            chunk = result['chunk']
+            chunk_metadata = chunk['chunkMetadata']
+
+            if 'previousChunks' in chunk_metadata:
+                # Process previous chunks
+                for prev_chunk in chunk['chunkMetadata']['previousChunks']:
+                    prev_chunk_string = (
+                        f"# {prev_chunk['id']}\n"
+                        f"{prev_chunk['content']}\n"
+                        f"## metadata\n"
+                        f"Document URI: {prev_chunk['documentMetadata']['uri']}\n"
+                        f"Document Title: {prev_chunk['documentMetadata']['title']}\n"
+                    )
+                    all_chunks.append(prev_chunk_string)
+
+            # Process fetched chunk
+            fetched_chunk_string = (
+                f"# {chunk['id']}\n"
+                f"{chunk['content']}\n"
+                f"## metadata\n"
+                f"Document URI: {chunk['documentMetadata']['uri']}\n"
+                f"Document Title: {chunk['documentMetadata']['title']}\n"
+            )
+            all_chunks.append(fetched_chunk_string)
+
+            # Process next chunks
+            if 'nextChunks' in chunk_metadata:
+                for next_chunk in chunk_metadata['nextChunks']:
+                    next_chunk_string = (
+                        f"# {next_chunk['id']}\n"
+                        f"{next_chunk['content']}\n"
+                        f"## metadata\n"
+                        f"Document URI: {next_chunk['documentMetadata']['uri']}\n"
+                        f"Document Title: {next_chunk['documentMetadata']['title']}\n"
+                    )
+                    all_chunks.append(next_chunk_string)
+
+        # Combine all chunks into one long string
+        result_string = "\n".join(all_chunks)
+
+        return result_string
+
     def import_documents(self,
         gcs_uri: Optional[str] = None,
         data_schema="content",
@@ -256,9 +317,9 @@ class DiscoveryEngineClient:
         try:
             operation = import_documents_with_retry(self.doc_client, request)
         except ResourceExhausted as e:
-            print(f"Operation failed after retries due to quota exceeded: {e}")
+            log.error(f"Operation failed after retries due to quota exceeded: {e}")
         except Exception as e:
-            print(f"An unexpected error occurred: {e}")
+            log.error(f"An unexpected error occurred: {e}")
 
         return operation.operation.name
 

{sunholo-0.69.9 → sunholo-0.69.12}/sunholo/llamaindex/import_files.py

@@ -61,7 +61,6 @@ def do_llamaindex(message_data, metadata, vector_name):
     corpuses = []
     for memory in memories:
         for key, value in memory.items():  # Now iterate over the dictionary
-            log.info(f"Found memory {key}")
             vectorstore = value.get('vectorstore')
             if vectorstore == "llamaindex":
                 log.info(f"Found vectorstore {vectorstore}")

{sunholo-0.69.9 → sunholo-0.69.12}/sunholo/logging.py

@@ -119,8 +119,14 @@ class GoogleCloudLogging:
         if log_text:
             if isinstance(log_struct, dict):
                 logger.log_struct(log_struct, severity=severity, source_location=caller_info)
-            else:
+            elif isinstance(log_struct, str):
                 logger.log_text(log_text, severity=severity, source_location=caller_info)
+            else:
+                try:
+                    turn_to_text = str(log_text)
+                    logger.log_text(turn_to_text, severity=severity, source_location=caller_info)
+                except Exception as err:
+                    print(f"Could not log this: {log_text=} - {str(err)}")
 
         elif log_struct:
             if not isinstance(log_struct, dict):

{sunholo-0.69.9 → sunholo-0.69.12}/sunholo.egg-info/PKG-INFO

@@ -1,9 +1,9 @@
 Metadata-Version: 2.1
 Name: sunholo
-Version: 0.69.9
+Version: 0.69.12
 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/v0.69.9.tar.gz
+Download-URL: https://github.com/sunholo-data/sunholo-py/archive/refs/tags/v0.69.12.tar.gz
 Author: Holosun ApS
 Author-email: multivac@sunholo.com
 License: Apache License, Version 2.0
@@ -61,6 +61,7 @@ Requires-Dist: pypdf; extra == "all"
 Requires-Dist: python-socketio; extra == "all"
 Requires-Dist: rich; extra == "all"
 Requires-Dist: supabase; extra == "all"
+Requires-Dist: tantivy; extra == "all"
 Requires-Dist: tiktoken; extra == "all"
 Provides-Extra: cli
 Requires-Dist: jsonschema>=4.21.1; extra == "cli"
@@ -73,6 +74,7 @@ Requires-Dist: pg8000; extra == "database"
 Requires-Dist: pgvector; extra == "database"
 Requires-Dist: psycopg2-binary; extra == "database"
 Requires-Dist: lancedb; extra == "database"
+Requires-Dist: tantivy; extra == "database"
 Provides-Extra: pipeline
 Requires-Dist: GitPython; extra == "pipeline"
 Requires-Dist: lark; extra == "pipeline"

{sunholo-0.69.9 → sunholo-0.69.12}/sunholo.egg-info/requires.txt

@@ -42,6 +42,7 @@ pypdf
 python-socketio
 rich
 supabase
+tantivy
 tiktoken
 
 [anthropic]
@@ -59,6 +60,7 @@ pg8000
 pgvector
 psycopg2-binary
 lancedb
+tantivy
 
 [gcp]
 google-auth-httplib2