vanna 0.7.7__tar.gz → 0.7.9__tar.gz

{vanna-0.7.7 → vanna-0.7.9}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: vanna
-Version: 0.7.7
+Version: 0.7.9
 Summary: Generate SQL queries from natural language
 Author-email: Zain Hoda <zain@vanna.ai>
 Requires-Python: >=3.9
@@ -28,7 +28,7 @@ Requires-Dist: duckdb ; extra == "all"
 Requires-Dist: openai ; extra == "all"
 Requires-Dist: qianfan ; extra == "all"
 Requires-Dist: mistralai>=1.0.0 ; extra == "all"
-Requires-Dist: chromadb ; extra == "all"
+Requires-Dist: chromadb<1.0.0 ; extra == "all"
 Requires-Dist: anthropic ; extra == "all"
 Requires-Dist: zhipuai ; extra == "all"
 Requires-Dist: marqo ; extra == "all"
@@ -64,7 +64,7 @@ Requires-Dist: fastembed ; extra == "azuresearch"
 Requires-Dist: boto3 ; extra == "bedrock"
 Requires-Dist: botocore ; extra == "bedrock"
 Requires-Dist: google-cloud-bigquery ; extra == "bigquery"
-Requires-Dist: chromadb ; extra == "chromadb"
+Requires-Dist: chromadb<1.0.0 ; extra == "chromadb"
 Requires-Dist: clickhouse_connect ; extra == "clickhouse"
 Requires-Dist: duckdb ; extra == "duckdb"
 Requires-Dist: faiss-cpu ; extra == "faiss-cpu"
@@ -85,7 +85,7 @@ Requires-Dist: opensearch-dsl ; extra == "opensearch"
 Requires-Dist: langchain-community ; extra == "opensearch"
 Requires-Dist: langchain-huggingface ; extra == "opensearch"
 Requires-Dist: oracledb ; extra == "oracle"
-Requires-Dist: chromadb ; extra == "oracle"
+Requires-Dist: chromadb<1.0.0 ; extra == "oracle"
 Requires-Dist: langchain-postgres>=0.0.12 ; extra == "pgvector"
 Requires-Dist: pinecone ; extra == "pinecone"
 Requires-Dist: fastembed ; extra == "pinecone"

{vanna-0.7.7 → vanna-0.7.9}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "flit_core.buildapi"
 
 [project]
 name = "vanna"
-version = "0.7.7"
+version = "0.7.9"
 authors = [
   { name="Zain Hoda", email="zain@vanna.ai" },
 ]
@@ -33,9 +33,9 @@ bigquery = ["google-cloud-bigquery"]
 snowflake = ["snowflake-connector-python"]
 duckdb = ["duckdb"]
 google = ["google-generativeai", "google-cloud-aiplatform"]
-all = ["psycopg2-binary", "db-dtypes", "PyMySQL", "google-cloud-bigquery", "snowflake-connector-python", "duckdb", "openai", "qianfan", "mistralai>=1.0.0", "chromadb", "anthropic", "zhipuai", "marqo", "google-generativeai", "google-cloud-aiplatform", "qdrant-client", "fastembed", "ollama", "httpx", "opensearch-py", "opensearch-dsl", "transformers", "pinecone", "pymilvus[model]","weaviate-client", "azure-search-documents", "azure-identity", "azure-common", "faiss-cpu", "boto", "boto3", "botocore", "langchain_core", "langchain_postgres", "langchain-community", "langchain-huggingface", "xinference-client"]
+all = ["psycopg2-binary", "db-dtypes", "PyMySQL", "google-cloud-bigquery", "snowflake-connector-python", "duckdb", "openai", "qianfan", "mistralai>=1.0.0", "chromadb<1.0.0", "anthropic", "zhipuai", "marqo", "google-generativeai", "google-cloud-aiplatform", "qdrant-client", "fastembed", "ollama", "httpx", "opensearch-py", "opensearch-dsl", "transformers", "pinecone", "pymilvus[model]","weaviate-client", "azure-search-documents", "azure-identity", "azure-common", "faiss-cpu", "boto", "boto3", "botocore", "langchain_core", "langchain_postgres", "langchain-community", "langchain-huggingface", "xinference-client"]
 test = ["tox"]
-chromadb = ["chromadb"]
+chromadb = ["chromadb<1.0.0"]
 openai = ["openai"]
 qianfan = ["qianfan"]
 mistralai = ["mistralai>=1.0.0"]
@@ -57,4 +57,4 @@ pgvector = ["langchain-postgres>=0.0.12"]
 faiss-cpu = ["faiss-cpu"]
 faiss-gpu = ["faiss-gpu"]
 xinference-client = ["xinference-client"]
-oracle = ["oracledb", "chromadb"]
+oracle = ["oracledb", "chromadb<1.0.0"]

{vanna-0.7.7 → vanna-0.7.9}/src/vanna/base/base.py

@@ -181,30 +181,47 @@ class VannaBase(ABC):
             str: The extracted SQL query.
         """
 
-        # If the llm_response contains a CTE (with clause), extract the last sql between WITH and ;
-        sqls = re.findall(r"\bWITH\b .*?;", llm_response, re.DOTALL)
+        import re
+        """
+        Extracts the SQL query from the LLM response, handling various formats including:
+        - WITH clause
+        - SELECT statement
+        - CREATE TABLE AS SELECT
+        - Markdown code blocks
+        """
+
+        # Match CREATE TABLE ... AS SELECT
+        sqls = re.findall(r"\bCREATE\s+TABLE\b.*?\bAS\b.*?;", llm_response, re.DOTALL | re.IGNORECASE)
         if sqls:
             sql = sqls[-1]
             self.log(title="Extracted SQL", message=f"{sql}")
             return sql
 
-        # If the llm_response is not markdown formatted, extract last sql by finding select and ; in the response
-        sqls = re.findall(r"SELECT.*?;", llm_response, re.DOTALL)
+        # Match WITH clause (CTEs)
+        sqls = re.findall(r"\bWITH\b .*?;", llm_response, re.DOTALL | re.IGNORECASE)
         if sqls:
             sql = sqls[-1]
             self.log(title="Extracted SQL", message=f"{sql}")
             return sql
 
-        # If the llm_response contains a markdown code block, with or without the sql tag, extract the last sql from it
-        sqls = re.findall(r"```sql\n(.*)```", llm_response, re.DOTALL)
+        # Match SELECT ... ;
+        sqls = re.findall(r"\bSELECT\b .*?;", llm_response, re.DOTALL | re.IGNORECASE)
         if sqls:
             sql = sqls[-1]
             self.log(title="Extracted SQL", message=f"{sql}")
             return sql
 
-        sqls = re.findall(r"```(.*)```", llm_response, re.DOTALL)
+        # Match ```sql ... ``` blocks
+        sqls = re.findall(r"```sql\s*\n(.*?)```", llm_response, re.DOTALL | re.IGNORECASE)
         if sqls:
-            sql = sqls[-1]
+            sql = sqls[-1].strip()
+            self.log(title="Extracted SQL", message=f"{sql}")
+            return sql
+
+        # Match any ``` ... ``` code blocks
+        sqls = re.findall(r"```(.*?)```", llm_response, re.DOTALL | re.IGNORECASE)
+        if sqls:
+            sql = sqls[-1].strip()
             self.log(title="Extracted SQL", message=f"{sql}")
             return sql
 

{vanna-0.7.7 → vanna-0.7.9}/src/vanna/pinecone/pinecone_vector.py

@@ -111,7 +111,7 @@ class PineconeDB_VectorStore(VannaBase):
 
     def _check_if_embedding_exists(self, id: str, namespace: str) -> bool:
         fetch_response = self.Index.fetch(ids=[id], namespace=namespace)
-        if fetch_response["vectors"] == {}:
+        if fetch_response.vectors == {}:
             return False
         return True