fix: Raise from exc related fixes, bumped to 0.25.1 (#47)

* Raise from exc related fixes, bumped to 0.25.1

* Index related error handling improv

Author: chandrasekharan-zipstack

src/unstract/sdk/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "0.25.0"
+__version__ = "0.25.1"
 
 
 def get_sdk_version():

src/unstract/sdk/embedding.py

@@ -45,7 +45,7 @@ def get_embedding(self, adapter_instance_id: str) -> BaseEmbedding:
             self.tool.stream_log(
                 log=f"Error getting embedding: {e}", level=LogLevel.ERROR
             )
-            raise ToolEmbeddingError(f"Error getting embedding instance: {e}")
+            raise ToolEmbeddingError(f"Error getting embedding instance: {e}") from e
 
     def get_embedding_length(self, embedding: BaseEmbedding) -> int:
         embedding_list = embedding._get_text_embedding(self.__TEST_SNIPPET)

src/unstract/sdk/exceptions.py

@@ -29,5 +29,9 @@ class ToolVectorDBError(SdkError):
     DEFAULT_MESSAGE = "Error ocurred related to vector DB"
 
 
+class X2TextError(SdkError):
+    DEFAULT_MESSAGE = "Error ocurred related to text extractor"
+
+
 class RateLimitError(SdkError):
     DEFAULT_MESSAGE = "Running into rate limit errors, please try again later"

src/unstract/sdk/index.py

@@ -196,7 +196,7 @@ def index_file(
                     f"Error deleting nodes for {doc_id}: {e}",
                     level=LogLevel.ERROR,
                 )
-                raise SdkError(f"Error deleting nodes for {doc_id}: {e}")
+                raise SdkError(f"Error deleting nodes for {doc_id}: {e}") from e
             doc_id_found = False
 
         if doc_id_found:
@@ -246,29 +246,33 @@ def index_file(
             document.id_ = doc_id
             documents.append(document)
         self.tool.stream_log(f"Number of documents: {len(documents)}")
-        if chunk_size == 0:
-            parser = SimpleNodeParser.from_defaults(
-                chunk_size=len(documents[0].text) + 10, chunk_overlap=0
-            )
-            nodes = parser.get_nodes_from_documents(documents, show_progress=True)
-            node = nodes[0]
-            node.embedding = embedding_li.get_query_embedding(" ")
-            vector_db_li.add(nodes=[node])
-            self.tool.stream_log("Added node to vector db")
-        else:
-            storage_context = StorageContext.from_defaults(vector_store=vector_db_li)
-            parser = SimpleNodeParser.from_defaults(
-                chunk_size=chunk_size, chunk_overlap=chunk_overlap
-            )
 
-            # Set callback_manager to collect Usage stats
-            callback_manager = UNCallbackManager.set_callback_manager(
-                platform_api_key=self.tool.get_env_or_die(ToolEnv.PLATFORM_API_KEY),
-                embedding=embedding_li,
-            )
+        try:
+            if chunk_size == 0:
+                parser = SimpleNodeParser.from_defaults(
+                    chunk_size=len(documents[0].text) + 10, chunk_overlap=0
+                )
+                nodes = parser.get_nodes_from_documents(documents, show_progress=True)
+                node = nodes[0]
+                node.embedding = embedding_li.get_query_embedding(" ")
+                vector_db_li.add(nodes=[node])
+                self.tool.stream_log("Added node to vector db")
+            else:
+                storage_context = StorageContext.from_defaults(
+                    vector_store=vector_db_li
+                )
+                parser = SimpleNodeParser.from_defaults(
+                    chunk_size=chunk_size, chunk_overlap=chunk_overlap
+                )
+
+                # Set callback_manager to collect Usage stats
+                callback_manager = UNCallbackManager.set_callback_manager(
+                    platform_api_key=self.tool.get_env_or_die(ToolEnv.PLATFORM_API_KEY),
+                    embedding=embedding_li,
+                )
+
+                self.tool.stream_log("Adding nodes to vector db...")
 
-            self.tool.stream_log("Adding nodes to vector db...")
-            try:
                 VectorStoreIndex.from_documents(
                     documents,
                     storage_context=storage_context,
@@ -277,13 +281,13 @@ def index_file(
                     node_parser=parser,
                     callback_manager=callback_manager,
                 )
-            except Exception as e:
-                self.tool.stream_log(
-                    f"Error adding nodes to vector db: {e}",
-                    level=LogLevel.ERROR,
-                )
-                raise IndexingError(str(e)) from e
-            self.tool.stream_log("Added nodes to vector db")
+        except Exception as e:
+            self.tool.stream_log(
+                f"Error adding nodes to vector db: {e}",
+                level=LogLevel.ERROR,
+            )
+            raise IndexingError(str(e)) from e
+        self.tool.stream_log("Added nodes to vector db")
 
         self.tool.stream_log("File has been indexed successfully")
         return doc_id

src/unstract/sdk/llm.py

@@ -103,7 +103,7 @@ def get_llm(self, adapter_instance_id: str) -> LLM:
             self.tool.stream_log(
                 log=f"Unable to get llm instance: {e}", level=LogLevel.ERROR
             )
-            raise ToolLLMError(f"Error getting llm instance: {e}")
+            raise ToolLLMError(f"Error getting llm instance: {e}") from e
 
     def get_max_tokens(self, reserved_for_output: int = 0) -> int:
         """Returns the maximum number of tokens that can be used for the LLM.

src/unstract/sdk/vector_db.py

@@ -77,4 +77,4 @@ def get_vector_db(
                 log=f"Unable to get vector_db {adapter_instance_id}: {e}",
                 level=LogLevel.ERROR,
             )
-            raise ToolVectorDBError(f"Error getting vectorDB instance: {e}")
+            raise ToolVectorDBError(f"Error getting vectorDB instance: {e}") from e

src/unstract/sdk/x2txt.py

@@ -7,7 +7,7 @@
 
 from unstract.sdk.adapters import ToolAdapter
 from unstract.sdk.constants import LogLevel
-from unstract.sdk.exceptions import SdkError
+from unstract.sdk.exceptions import X2TextError
 from unstract.sdk.tool.base import BaseTool
 
 
@@ -28,17 +28,15 @@ def get_x2text(self, adapter_instance_id: str) -> X2TextAdapter:
                 ][Common.ADAPTER]
                 x2text_metadata = x2text_config.get(Common.ADAPTER_METADATA)
                 # Add x2text service host, port and platform_service_key
-                x2text_metadata[
+                x2text_metadata[X2TextConstants.X2TEXT_HOST] = self.tool.get_env_or_die(
                     X2TextConstants.X2TEXT_HOST
-                ] = self.tool.get_env_or_die(X2TextConstants.X2TEXT_HOST)
-                x2text_metadata[
+                )
+                x2text_metadata[X2TextConstants.X2TEXT_PORT] = self.tool.get_env_or_die(
                     X2TextConstants.X2TEXT_PORT
-                ] = self.tool.get_env_or_die(X2TextConstants.X2TEXT_PORT)
+                )
                 x2text_metadata[
                     X2TextConstants.PLATFORM_SERVICE_API_KEY
-                ] = self.tool.get_env_or_die(
-                    X2TextConstants.PLATFORM_SERVICE_API_KEY
-                )
+                ] = self.tool.get_env_or_die(X2TextConstants.PLATFORM_SERVICE_API_KEY)
 
                 x2text_adapter_class = x2text_adapter(x2text_metadata)
 
@@ -49,4 +47,4 @@ def get_x2text(self, adapter_instance_id: str) -> X2TextAdapter:
                 log=f"Unable to get x2text adapter {adapter_instance_id}: {e}",
                 level=LogLevel.ERROR,
             )
-            raise SdkError(f"Error getting vectorDB instance: {e}")
+            raise X2TextError(f"Error getting text extractor: {e}") from e