feat: Improved error handling for index, bumped SDK to 0.15.0

pyproject.toml

@@ -15,7 +15,7 @@ dependencies = [
     "python-magic~=0.4.27",
     "python-dotenv==1.0.0",
     # LLM Triad
-    "unstract-adapters~=0.4.0",
+    "unstract-adapters~=0.4.1",
     "llama-index==0.9.28",
     "tiktoken~=0.4.0",
     "transformers==4.37.0",

src/unstract/sdk/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "0.14.0"
+__version__ = "0.15.0"
 
 
 def get_sdk_version():

src/unstract/sdk/exceptions.py

@@ -1,13 +1,10 @@
-from typing import Any, Optional
+class SdkError(Exception):
+    DEFAULT_MESSAGE = "Something went wrong"
 
+    def __init__(self, message: str = DEFAULT_MESSAGE):
+        super().__init__(message)
+        # Make it user friendly wherever possible
+        self.message = message
 
-class SdkException(Exception):
-    def __init__(
-        self, *args: Any, user_message: Optional[str] = None, **kwargs: Any
-    ) -> None:
-        super().__init__(*args, **kwargs)
-        self._user_message = user_message
-
-    @property
-    def user_message(self) -> Optional[str]:
-        return self._user_message
+    def __str__(self) -> str:
+        return self.message

src/unstract/sdk/index.py

@@ -3,11 +3,12 @@
 from llama_index import Document, StorageContext, VectorStoreIndex
 from llama_index.node_parser import SimpleNodeParser
 from llama_index.vector_stores import VectorStoreQuery, VectorStoreQueryResult
+from unstract.adapters.exceptions import AdapterError
 from unstract.adapters.x2text.x2text_adapter import X2TextAdapter
 
 from unstract.sdk.constants import LogLevel, ToolEnv
 from unstract.sdk.embedding import ToolEmbedding
-from unstract.sdk.exceptions import SdkException
+from unstract.sdk.exceptions import SdkError
 from unstract.sdk.tool.base import BaseTool
 from unstract.sdk.utils import ToolUtils
 from unstract.sdk.utils.service_context import ServiceContext
@@ -30,7 +31,7 @@ def get_text_from_index(
             self.tool.stream_log(
                 f"Error loading {embedding_type}", level=LogLevel.ERROR
             )
-            raise SdkException(f"Error loading {embedding_type}")
+            raise SdkError(f"Error loading {embedding_type}")
         embedding_dimension = embedd_helper.get_embedding_length(embedding_li)
 
         vdb_helper = ToolVectorDB(
@@ -45,7 +46,7 @@ def get_text_from_index(
             self.tool.stream_log(
                 f"Error loading {vector_db}", level=LogLevel.ERROR
             )
-            raise SdkException(f"Error loading {vector_db}")
+            raise SdkError(f"Error loading {vector_db}")
 
         try:
             self.tool.stream_log(f">>> Querying {vector_db}...")
@@ -59,7 +60,7 @@ def get_text_from_index(
             self.tool.stream_log(
                 f"Error querying {vector_db}: {e}", level=LogLevel.ERROR
             )
-            raise SdkException(f"Error querying {vector_db}: {e}")
+            raise SdkError(f"Error querying {vector_db}: {e}")
 
         n: VectorStoreQueryResult = vector_db_li.query(query=q)
         if len(n.nodes) > 0:
@@ -134,13 +135,18 @@ def index_file(
 
         self.tool.stream_log("Extracting text from input file")
         full_text = []
-        x2text = X2Text(tool=self.tool)
-        x2text_adapter_inst: X2TextAdapter = x2text.get_x2text(
-            adapter_instance_id=x2text_adapter
-        )
-        extracted_text = x2text_adapter_inst.process(
-            input_file_path=file_path, output_file_path=output_file_path
-        )
+        extracted_text = ""
+        try:
+            x2text = X2Text(tool=self.tool)
+            x2text_adapter_inst: X2TextAdapter = x2text.get_x2text(
+                adapter_instance_id=x2text_adapter
+            )
+            extracted_text = x2text_adapter_inst.process(
+                input_file_path=file_path, output_file_path=output_file_path
+            )
+        except AdapterError as e:
+            # Wrapping AdapterErrors with SdkError
+            raise SdkError(str(e)) from e
         full_text.append(
             {
                 "section": "full",
@@ -173,7 +179,7 @@ def index_file(
             self.tool.stream_log(
                 f"Error loading {embedding_type}", level=LogLevel.ERROR
             )
-            raise SdkException(f"Error loading {embedding_type}")
+            raise SdkError(f"Error loading {embedding_type}")
 
         embedding_dimension = embedd_helper.get_embedding_length(embedding_li)
         vector_db_li = vdb_helper.get_vector_db(
@@ -184,7 +190,7 @@ def index_file(
             self.tool.stream_log(
                 f"Error loading {vector_db}", level=LogLevel.ERROR
             )
-            raise SdkException(f"Error loading {vector_db}")
+            raise SdkError(f"Error loading {vector_db}")
 
         q = VectorStoreQuery(
             query_embedding=embedding_li.get_query_embedding(" "),
@@ -214,7 +220,7 @@ def index_file(
                     f"Error deleting nodes for {doc_id}: {e}",
                     level=LogLevel.ERROR,
                 )
-                raise SdkException(f"Error deleting nodes for {doc_id}: {e}")
+                raise SdkError(f"Error deleting nodes for {doc_id}: {e}")
             doc_id_not_found = True
 
         if doc_id_not_found:
@@ -271,7 +277,7 @@ def index_file(
                         f"Error adding nodes to vector db: {e}",
                         level=LogLevel.ERROR,
                     )
-                    raise SdkException(f"Error adding nodes to vector db: {e}")
+                    raise SdkError(f"Error adding nodes to vector db: {e}")
                 self.tool.stream_log("Added nodes to vector db")
 
         self.tool.stream_log("Done indexing file")

src/unstract/sdk/vector_db.py

@@ -5,9 +5,10 @@
 from unstract.adapters.constants import Common
 from unstract.adapters.vectordb import adapters
 from unstract.adapters.vectordb.constants import VectorDbConstants
+
 from unstract.sdk.adapters import ToolAdapter
 from unstract.sdk.constants import LogLevel, ToolEnv, ToolSettingsKey
-from unstract.sdk.exceptions import SdkException
+from unstract.sdk.exceptions import SdkError
 from unstract.sdk.platform import PlatformHelper
 from unstract.sdk.tool.base import BaseTool
 
@@ -34,7 +35,7 @@ def __get_org_id(self) -> str:
         platform_details = platform_helper.get_platform_details()
         if not platform_details:
             # Errors are logged by the SDK itself
-            raise SdkException("Error getting platform details")
+            raise SdkError("Error getting platform details")
         account_id = platform_details.get("organization_id")
         return account_id