wizit-context-ingestor 0.3.0b4__py3-none-any.whl → 0.3.0b6__py3-none-any.whl

wizit_context_ingestor/application/transcription_service.py

@@ -112,7 +112,7 @@ class TranscriptionService:
     #         raise
 
     async def parse_doc_page_with_workflow(
-        self, document: ParsedDocPage
+        self, document: ParsedDocPage, retries: int = 0
     ) -> ParsedDocPage:
         """Transcribe an image to text using an agent.
         Args:
@@ -120,6 +120,9 @@ class TranscriptionService:
         Returns:
             Processed text
         """
+        if retries > 1:
+            logger.info("Max retries exceeded")
+            return document
         result = await self.compiled_transcription_workflow.ainvoke(
             {
                 "messages": [
@@ -150,10 +153,12 @@ class TranscriptionService:
                 }
             },
         )
-        if result["transcription"]:
+        if "transcription" in result:
             document.page_text = result["transcription"]
         else:
-            raise ValueError(f"No transcription found: {result} ")
+            return await self.parse_doc_page_with_workflow(
+                document, retries=retries + 1
+            )
         return document
 
     # def process_document(self, file_key: str) -> Tuple[List[ParsedDocPage], ParsedDoc]:

wizit_context_ingestor/utils/file_utils.py

@@ -7,7 +7,7 @@ def has_invalid_file_name_format(file_name):
     if " " in file_name:
         return True
 
-    # Check for special characters (anything that's not alphanumeric, underscore, dash, or dot)
-    if re.search(r"[^a-zA-Z0-9_.-]", file_name):
+    # Check for special characters (anything that's not alphanumeric, underscore, dash, dot, slash, or backslash)
+    if re.search(r"[^a-zA-Z0-9_.-/\\]", file_name):
         return True
     return False

{wizit_context_ingestor-0.3.0b4.dist-info → wizit_context_ingestor-0.3.0b6.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: wizit-context-ingestor
-Version: 0.3.0b4
+Version: 0.3.0b6
 Summary: Contextual Rag with Cloud Solutions
 Requires-Dist: anthropic[vertex]>=0.66.0
 Requires-Dist: boto3>=1.40.23

{wizit_context_ingestor-0.3.0b4.dist-info → wizit_context_ingestor-0.3.0b6.dist-info}/RECORD

@@ -3,7 +3,7 @@ wizit_context_ingestor/__init__.py,sha256=TSTm5qSpNNCz9ilKYkXRUxupvmWG2AHfv7RBWF
 wizit_context_ingestor/application/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 wizit_context_ingestor/application/context_chunk_service.py,sha256=LYRKBsY30IT2LczkgkYdPx7W3yycRy-0m7t3KKgq6Nw,9046
 wizit_context_ingestor/application/interfaces.py,sha256=W0qonE3t-S-zwAoKtDYc4oyW_GOILKVmrdy8LnC8MVI,3193
-wizit_context_ingestor/application/transcription_service.py,sha256=jAjQE_sR0E3CSHLf0lq-24scl-_VKWy3crGhiodkoSM,8394
+wizit_context_ingestor/application/transcription_service.py,sha256=FlUcMGyAotAO8MmT5UMlPMbgIWVQLg7YO6rJx9ANn7A,8567
 wizit_context_ingestor/data/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 wizit_context_ingestor/data/kdb.py,sha256=GCkXQmnk2JCXV_VJ-h0k55AOIX8qohzBJN2v-9D1dlU,194
 wizit_context_ingestor/data/prompts.py,sha256=EnocoriDjPcFPd6Af9G6TUTB8NkO4EFN4AUHfpRVqYU,14406
@@ -29,7 +29,7 @@ wizit_context_ingestor/services/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm
 wizit_context_ingestor/services/chunks.py,sha256=tQQsdsOscZWzqVY5WxVxr3ii62FOJ3nMARaJJz6CvjQ,2011
 wizit_context_ingestor/services/parse_doc.py,sha256=3CyZoGbiUfxbs0SXUWXjQevtusSzTBgvUVeNNSdxJLE,4491
 wizit_context_ingestor/services/pg_embeddings_manager.py,sha256=n1HOmu_Z_Z71H-rVAyJS3FdPKbBckm5W8_XethY8nuM,4998
-wizit_context_ingestor/utils/file_utils.py,sha256=QnyncN0X5E-LjAYxFPxQiOrAj0DHcAcL2GliLVikF5o,393
+wizit_context_ingestor/utils/file_utils.py,sha256=Wuua14LivrfL8oBP-j3ZtSqc7uq4rrQzT-T-p7pxEpM,414
 wizit_context_ingestor/workflows/context_nodes.py,sha256=3qlFcxPUmehx04mQHpmouneKq--To8rwSDHCRFyWICo,3168
 wizit_context_ingestor/workflows/context_state.py,sha256=4MTIUjK-F2pWvIldovWZhMAqqCOpViKbvitJzETkSkY,324
 wizit_context_ingestor/workflows/context_tools.py,sha256=E9VTL3AC0MwSIuc1e-juZK7XCxnZfFv0-KpHfR2CNH4,2764
@@ -39,6 +39,6 @@ wizit_context_ingestor/workflows/transcription_schemas.py,sha256=CQCl7LXD5voxhJO
 wizit_context_ingestor/workflows/transcription_state.py,sha256=2Z_t2aZFEH_nAjdEO6RFBEmi_fwvr9cV0aLS1eIxiCQ,590
 wizit_context_ingestor/workflows/transcription_tools.py,sha256=FtIfWFITn8_Rr5SEobCeR55aJGZoHRMgF2UxRT5vJ-E,1373
 wizit_context_ingestor/workflows/transcription_workflow.py,sha256=77cLsYGdv01Py2GaKYpACuifPeSxH7tkVodvLv97sdg,1621
-wizit_context_ingestor-0.3.0b4.dist-info/WHEEL,sha256=eh7sammvW2TypMMMGKgsM83HyA_3qQ5Lgg3ynoecH3M,79
-wizit_context_ingestor-0.3.0b4.dist-info/METADATA,sha256=iPcDYUP3VQKukxafk_9HeJQqzw8WqLKrZH71cmbuIYw,3768
-wizit_context_ingestor-0.3.0b4.dist-info/RECORD,,
+wizit_context_ingestor-0.3.0b6.dist-info/WHEEL,sha256=eh7sammvW2TypMMMGKgsM83HyA_3qQ5Lgg3ynoecH3M,79
+wizit_context_ingestor-0.3.0b6.dist-info/METADATA,sha256=zp_AzRkME9KJkF_9yyiVLE2dNwxPxROK9bmivV8tOMw,3768
+wizit_context_ingestor-0.3.0b6.dist-info/RECORD,,