unstructured-ingest 1.0.37__tar.gz → 1.0.40__tar.gz

{unstructured_ingest-1.0.37 → unstructured_ingest-1.0.40}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: unstructured_ingest
-Version: 1.0.37
+Version: 1.0.40
 Summary: Local ETL data pipeline to get data RAG ready
 Author-email: Unstructured Technologies <devops@unstructuredai.io>
 License-Expression: Apache-2.0

unstructured_ingest-1.0.40/unstructured_ingest/__version__.py

@@ -0,0 +1 @@
+__version__ = "1.0.40"  # pragma: no cover

{unstructured_ingest-1.0.37 → unstructured_ingest-1.0.40}/unstructured_ingest/processes/connectors/confluence.py

@@ -33,6 +33,8 @@ from unstructured_ingest.utils.string_and_date_utils import fix_unescaped_unicod
 
 if TYPE_CHECKING:
     from atlassian import Confluence
+    from bs4 import BeautifulSoup
+    from bs4.element import Tag
 
 CONNECTOR_TYPE = "confluence"
 
@@ -235,11 +237,28 @@ class ConfluenceIndexer(Indexer):
                 yield file_data
 
 
-class ConfluenceDownloaderConfig(DownloaderConfig, HtmlMixin):
+class ConfluenceDownloaderConfig(HtmlMixin, DownloaderConfig):
     max_num_metadata_permissions: int = Field(
         250, description="Approximate maximum number of permissions included in metadata"
     )
 
+    @requires_dependencies(["bs4"])
+    def _find_hyperlink_tags(self, html_soup: "BeautifulSoup") -> list["Tag"]:
+        from bs4.element import Tag
+
+        return [
+            element
+            for element in html_soup.find_all(
+                "a",
+                attrs={
+                    "class": "confluence-embedded-file",
+                    "data-linked-resource-type": "attachment",
+                    "href": True,
+                },
+            )
+            if isinstance(element, Tag)
+        ]
+
 
 @dataclass
 class ConfluenceDownloader(Downloader):

{unstructured_ingest-1.0.37 → unstructured_ingest-1.0.40}/unstructured_ingest/processes/connectors/milvus.py

@@ -1,7 +1,7 @@
 import json
 from contextlib import contextmanager
 from dataclasses import dataclass, field
-from typing import TYPE_CHECKING, Any, Generator, Optional, Union
+from typing import TYPE_CHECKING, Any, Generator, Optional
 
 from dateutil import parser
 from pydantic import Field, Secret
@@ -97,10 +97,16 @@ class MilvusUploadStager(UploadStager):
 
     def conform_dict(self, element_dict: dict, file_data: FileData) -> dict:
         working_data = element_dict.copy()
-        if self.upload_stager_config.flatten_metadata and (
-            metadata := working_data.pop("metadata", None)
-        ):
-            working_data.update(flatten_dict(metadata, keys_to_omit=["data_source_record_locator"]))
+
+        if self.upload_stager_config.flatten_metadata:
+            metadata: dict[str, Any] = working_data.pop("metadata", {})
+            flattened_metadata = flatten_dict(
+                metadata,
+                separator="_",
+                flatten_lists=False,
+                remove_none=True,
+            )
+            working_data.update(flattened_metadata)
 
         # TODO: milvus sdk doesn't seem to support defaults via the schema yet,
         #  remove once that gets updated
@@ -154,6 +160,23 @@ class MilvusUploader(Uploader):
     upload_config: MilvusUploaderConfig
     connector_type: str = CONNECTOR_TYPE
 
+    def has_dynamic_fields_enabled(self) -> bool:
+        """Check if the target collection has dynamic fields enabled."""
+        try:
+            with self.get_client() as client:
+                collection_info = client.describe_collection(self.upload_config.collection_name)
+
+                # Check if dynamic field is enabled
+                # The schema info should contain enable_dynamic_field or enableDynamicField
+                schema_info = collection_info.get(
+                    "enable_dynamic_field",
+                    collection_info.get("enableDynamicField", False),
+                )
+                return bool(schema_info)
+        except Exception as e:
+            logger.warning(f"Could not determine if collection has dynamic fields enabled: {e}")
+            return False
+
     @DestinationConnectionError.wrap
     def precheck(self):
         from pymilvus import MilvusException
@@ -164,6 +187,7 @@ class MilvusUploader(Uploader):
                     raise DestinationConnectionError(
                         f"Collection '{self.upload_config.collection_name}' does not exist"
                     )
+
         except MilvusException as milvus_exception:
             raise DestinationConnectionError(
                 f"failed to precheck Milvus: {str(milvus_exception.message)}"
@@ -193,16 +217,66 @@ class MilvusUploader(Uploader):
             )
 
     @requires_dependencies(["pymilvus"], extras="milvus")
-    def insert_results(self, data: Union[dict, list[dict]]):
+    def _prepare_data_for_insert(self, data: list[dict]) -> list[dict]:
+        """
+        Conforms the provided data to the schema of the target Milvus collection.
+        - If dynamic fields are enabled, it ensures JSON-stringified fields are decoded.
+        - If dynamic fields are disabled, it filters out any fields not present in the schema.
+        """
+
+        dynamic_fields_enabled = self.has_dynamic_fields_enabled()
+
+        # If dynamic fields are enabled, 'languages' field needs to be a list
+        if dynamic_fields_enabled:
+            logger.debug("Dynamic fields enabled, ensuring 'languages' field is a list.")
+            prepared_data = []
+            for item in data:
+                new_item = item.copy()
+                if "languages" in new_item and isinstance(new_item["languages"], str):
+                    try:
+                        new_item["languages"] = json.loads(new_item["languages"])
+                    except (json.JSONDecodeError, TypeError):
+                        logger.warning(
+                            f"Could not JSON decode languages field: {new_item['languages']}. "
+                            "Leaving as string.",
+                        )
+                prepared_data.append(new_item)
+            return prepared_data
+
+        # If dynamic fields are not enabled, we need to filter out the metadata fields
+        # to avoid insertion errors for fields not defined in the schema
+        with self.get_client() as client:
+            collection_info = client.describe_collection(
+                self.upload_config.collection_name,
+            )
+        schema_fields = {
+            field["name"]
+            for field in collection_info.get("fields", [])
+            if not field.get("auto_id", False)
+        }
+        # Remove metadata fields that are not part of the base schema
+        filtered_data = []
+        for item in data:
+            filtered_item = {key: value for key, value in item.items() if key in schema_fields}
+            filtered_data.append(filtered_item)
+        return filtered_data
+
+    @requires_dependencies(["pymilvus"], extras="milvus")
+    def insert_results(self, data: list[dict]):
         from pymilvus import MilvusException
 
         logger.info(
             f"uploading {len(data)} entries to {self.connection_config.db_name} "
             f"db in collection {self.upload_config.collection_name}"
         )
+
+        prepared_data = self._prepare_data_for_insert(data=data)
+
         with self.get_client() as client:
             try:
-                res = client.insert(collection_name=self.upload_config.collection_name, data=data)
+                res = client.insert(
+                    collection_name=self.upload_config.collection_name, data=prepared_data
+                )
             except MilvusException as milvus_exception:
                 raise WriteError(
                     f"failed to upload records to Milvus: {str(milvus_exception.message)}"

{unstructured_ingest-1.0.37 → unstructured_ingest-1.0.40}/unstructured_ingest/processes/connectors/notion/types/blocks/synced_block.py

@@ -19,11 +19,11 @@ class OriginalSyncedBlock(BlockBase):
     @classmethod
     def from_dict(cls, data: dict):
         """Create OriginalSyncedBlock from dictionary data.
-        
+
         Original blocks contain children content.
         """
         if "children" not in data:
-             raise ValueError(f"OriginalSyncedBlock data missing 'children': {data}")
+            raise ValueError(f"OriginalSyncedBlock data missing 'children': {data}")
         return cls(children=data["children"])
 
     def get_html(self) -> Optional[HtmlTag]:
@@ -38,7 +38,7 @@ class DuplicateSyncedBlock(BlockBase):
     @staticmethod
     def can_have_children() -> bool:
         """Check if duplicate synced blocks can have children.
-        
+
         Duplicate blocks themselves don't have children directly fetched here,
         but they represent content that does, so Notion API might report has_children=True
         on the parent block object. The actual children are fetched from the original block.
@@ -48,7 +48,7 @@ class DuplicateSyncedBlock(BlockBase):
     @classmethod
     def from_dict(cls, data: dict):
         """Create DuplicateSyncedBlock from dictionary data.
-        
+
         Duplicate blocks contain a 'synced_from' reference.
         """
         synced_from_data = data.get("synced_from")
@@ -63,7 +63,7 @@ class DuplicateSyncedBlock(BlockBase):
 
     def get_html(self) -> Optional[HtmlTag]:
         """Get HTML representation of the duplicate synced block.
-        
+
         HTML representation might need fetching the original block's content,
         which is outside the scope of this simple data class.
         """
@@ -74,7 +74,7 @@ class SyncBlock(BlockBase):
     @staticmethod
     def can_have_children() -> bool:
         """Check if synced blocks can have children.
-        
+
         Synced blocks (both original and duplicate) can conceptually have children.
         """
         return True
@@ -82,7 +82,7 @@ class SyncBlock(BlockBase):
     @classmethod
     def from_dict(cls, data: dict):
         """Create appropriate SyncedBlock subclass from dictionary data.
-        
+
         Determine if it's a duplicate (has 'synced_from') or original (has 'children').
         """
         if data.get("synced_from") is not None:
@@ -99,10 +99,9 @@ class SyncBlock(BlockBase):
             # Consider logging a warning here if strictness is needed.
             return OriginalSyncedBlock(children=[])
 
-
     def get_html(self) -> Optional[HtmlTag]:
         """Get HTML representation of the synced block.
-        
+
         The specific instance returned by from_dict (Original or Duplicate)
         will handle its own get_html logic.
         This method on the base SyncBlock might not be directly called.

{unstructured_ingest-1.0.37 → unstructured_ingest-1.0.40}/unstructured_ingest/processes/connectors/pinecone.py

@@ -240,7 +240,7 @@ class PineconeUploader(VectorDBUploader):
         destination_name: str = "unstructuredautocreated",
         destination_type: Literal["pod", "serverless"] = "serverless",
         serverless_cloud: str = "aws",
-        serverless_region: str = "us-west-2",
+        serverless_region: str = "us-east-1",
         pod_environment: str = "us-east1-gcp",
         pod_type: str = "p1.x1",
         pod_count: int = 1,

{unstructured_ingest-1.0.37 → unstructured_ingest-1.0.40}/unstructured_ingest/utils/html.py

@@ -12,6 +12,7 @@ from unstructured_ingest.logger import logger
 from unstructured_ingest.utils.dep_check import requires_dependencies
 
 if TYPE_CHECKING:
+    from bs4 import BeautifulSoup
     from bs4.element import Tag
     from requests import Session
 
@@ -96,7 +97,7 @@ class HtmlMixin(BaseModel):
         from bs4 import BeautifulSoup
 
         soup = BeautifulSoup(html, "html.parser")
-        tags = soup.find_all("a", href=True)
+        tags = self._find_hyperlink_tags(soup)
         hrefs = [
             tag["href"]
             for tag in tags
@@ -158,3 +159,15 @@ class HtmlMixin(BaseModel):
             )
             for url_to_download in urls_to_download
         ]
+
+    @requires_dependencies(["bs4"])
+    def _find_hyperlink_tags(self, html_soup: "BeautifulSoup") -> list["Tag"]:
+        """Find hyperlink tags in the HTML.
+
+        Overwrite this method to customize the tag search.
+        """
+        from bs4.element import Tag
+
+        return [
+            element for element in html_soup.find_all("a", href=True) if isinstance(element, Tag)
+        ]

unstructured_ingest-1.0.37/unstructured_ingest/__version__.py

@@ -1 +0,0 @@
-__version__ = "1.0.37"  # pragma: no cover