rename complete

MANIFEST.in

@@ -17,7 +17,7 @@ include requirements/huggingface.in
 
 # Ingest extras
 include requirements/ingest/airtable.in
-include requirements/ingest/astra.in
+include requirements/ingest/astradb.in
 include requirements/ingest/azure-cognitive-search.in
 include requirements/ingest/azure.in
 include requirements/ingest/biomed.in

Makefile

@@ -253,9 +253,9 @@ install-ingest-mongodb:
 install-ingest-databricks-volumes:
 	python3 -m pip install -r requirements/ingest/databricks-volumes.txt
 
-.PHONY: install-ingest-astra
-install-ingest-astra:
-	python3 -m pip install -r requirements/ingest/astra.txt
+.PHONY: install-ingest-astradb
+install-ingest-astradb:
+	python3 -m pip install -r requirements/ingest/astradb.txt
 
 .PHONY: install-ingest-clarifai
 install-ingest-clarifai:

requirements/ingest/astra.txt → requirements/ingest/astradb.txt

@@ -2,15 +2,15 @@
 # This file is autogenerated by pip-compile with Python 3.9
 # by the following command:
 #
-#    pip-compile ./ingest/astra.in
+#    pip-compile ./ingest/astradb.in
 #
 anyio==3.7.1
     # via
     #   -c ./ingest/../base.txt
     #   -c ./ingest/../deps/constraints.txt
     #   httpx
 astrapy==1.4.0
-    # via -r ./ingest/astra.in
+    # via -r ./ingest/astradb.in
 bson==0.5.10
     # via astrapy
 cassandra-driver==3.29.1

setup.py

@@ -129,7 +129,7 @@ def load_requirements(file_list: Optional[Union[str, List[str]]] = None) -> List
         "xlsx": xlsx_reqs,
         # Extra requirements for data connectors
         "airtable": load_requirements("requirements/ingest/airtable.in"),
-        "astra": load_requirements("requirements/ingest/astra.in"),
+        "astradb": load_requirements("requirements/ingest/astradb.in"),
         "azure": load_requirements("requirements/ingest/azure.in"),
         "azure-cognitive-search": load_requirements(
             "requirements/ingest/azure-cognitive-search.in",

test_unstructured_ingest/dest/astra.sh → test_unstructured_ingest/dest/astradb.sh

@@ -5,7 +5,7 @@ set -e
 SRC_PATH=$(dirname "$(realpath "$0")")
 SCRIPT_DIR=$(dirname "$SRC_PATH")
 cd "$SCRIPT_DIR"/.. || exit 1
-OUTPUT_FOLDER_NAME=astra-dest
+OUTPUT_FOLDER_NAME=astradb-dest
 OUTPUT_DIR=$SCRIPT_DIR/structured-output/$OUTPUT_FOLDER_NAME
 WORK_DIR=$SCRIPT_DIR/workdir/$OUTPUT_FOLDER_NAME
 max_processes=${MAX_PROCESSES:=$(python3 -c "import os; print(os.cpu_count())")}
@@ -21,7 +21,7 @@ if [ -z "$ASTRA_DB_API_ENDPOINT" ]; then
 fi
 
 RANDOM_SUFFIX=$((RANDOM % 100000 + 1))
-COLLECTION_NAME="astra_test_output_$RANDOM_SUFFIX"
+COLLECTION_NAME="astradb_test_output_$RANDOM_SUFFIX"
 EMBEDDING_DIMENSION=384
 
 # shellcheck disable=SC1091
@@ -31,7 +31,7 @@ function cleanup() {
   cleanup_dir "$OUTPUT_DIR"
   cleanup_dir "$WORK_DIR"
 
-  python "$SCRIPT_DIR"/python/test-ingest-astra-output.py \
+  python "$SCRIPT_DIR"/python/test-ingest-astradb-output.py \
     --token "$ASTRA_DB_APPLICATION_TOKEN" \
     --api-endpoint "$ASTRA_DB_API_ENDPOINT" \
     --collection-name "$COLLECTION_NAME" down
@@ -51,14 +51,14 @@ PYTHONPATH=. ./unstructured/ingest/main.py \
   --chunk-max-characters 1500 \
   --chunk-multipage-sections \
   --embedding-provider "langchain-huggingface" \
-  astra \
+  astradb \
   --token "$ASTRA_DB_APPLICATION_TOKEN" \
   --api-endpoint "$ASTRA_DB_API_ENDPOINT" \
   --collection-name "$COLLECTION_NAME" \
   --embedding-dimension "$EMBEDDING_DIMENSION" \
   --requested-indexing-policy '{"deny": ["metadata"]}'
 
-python "$SCRIPT_DIR"/python/test-ingest-astra-output.py \
+python "$SCRIPT_DIR"/python/test-ingest-astradb-output.py \
   --token "$ASTRA_DB_APPLICATION_TOKEN" \
   --api-endpoint "$ASTRA_DB_API_ENDPOINT" \
   --collection-name "$COLLECTION_NAME" check

test_unstructured_ingest/python/test-ingest-astra-output.py → test_unstructured_ingest/python/test-ingest-astradb-output.py

@@ -10,7 +10,7 @@ def get_client(token, api_endpoint, collection_name) -> AstraDB:
     return astra_db, astra_db_collection
 
 
-@click.group(name="astra-ingest")
+@click.group(name="astradb-ingest")
 @click.option("--token", type=str)
 @click.option("--api-endpoint", type=str)
 @click.option("--collection-name", type=str, default="collection_test")

test_unstructured_ingest/src/astra.sh → test_unstructured_ingest/src/astradb.sh

@@ -5,7 +5,7 @@ set -e
 SRC_PATH=$(dirname "$(realpath "$0")")
 SCRIPT_DIR=$(dirname "$SRC_PATH")
 cd "$SCRIPT_DIR"/.. || exit 1
-OUTPUT_FOLDER_NAME=astra
+OUTPUT_FOLDER_NAME=astradb
 OUTPUT_DIR=$SCRIPT_DIR/structured-output/$OUTPUT_FOLDER_NAME
 WORK_DIR=$SCRIPT_DIR/workdir/$OUTPUT_FOLDER_NAME
 DOWNLOAD_DIR=$SCRIPT_DIR/download/$OUTPUT_FOLDER_NAME
@@ -23,7 +23,7 @@ fi
 COLLECTION_NAME="ingest_test_src"
 
 PYTHONPATH=. ./unstructured/ingest/main.py \
-  astra \
+  astradb \
   --token "$ASTRA_DB_APPLICATION_TOKEN" \
   --api-endpoint "$ASTRA_DB_API_ENDPOINT" \
   --collection-name "$COLLECTION_NAME" \

test_unstructured_ingest/test-ingest-dest.sh

@@ -15,7 +15,7 @@ cd "$SCRIPT_DIR"/.. || exit 1
 export OMP_THREAD_LIMIT=1
 
 all_tests=(
-  'astra.sh'
+  'astradb.sh'
   'azure.sh'
   'azure-cognitive-search.sh'
   'box.sh'

test_unstructured_ingest/test-ingest-src.sh

@@ -19,7 +19,7 @@ export OMP_THREAD_LIMIT=1
 all_tests=(
   's3.sh'
   's3-minio.sh'
-  'astra.sh'
+  'astradb.sh'
   'azure.sh'
   'biomed-api.sh'
   'biomed-path.sh'

unstructured/ingest/cli/cmds/__init__.py

@@ -7,8 +7,8 @@
 from unstructured.ingest.cli.cmds.fsspec.sftp import get_base_src_cmd as sftp_base_src_cmd
 
 from .airtable import get_base_src_cmd as airtable_base_src_cmd
-from .astra import get_base_dest_cmd as astra_base_dest_cmd
-from .astra import get_base_src_cmd as astra_base_src_cmd
+from .astradb import get_base_dest_cmd as astradb_base_dest_cmd
+from .astradb import get_base_src_cmd as astradb_base_src_cmd
 from .azure_cognitive_search import get_base_dest_cmd as azure_cognitive_search_base_dest_cmd
 from .biomed import get_base_src_cmd as biomed_base_src_cmd
 from .chroma import get_base_dest_cmd as chroma_base_dest_cmd
@@ -63,7 +63,7 @@
 
 base_src_cmd_fns: t.List[t.Callable[[], BaseSrcCmd]] = [
     airtable_base_src_cmd,
-    astra_base_src_cmd,
+    astradb_base_src_cmd,
     azure_base_src_cmd,
     biomed_base_src_cmd,
     box_base_src_cmd,
@@ -106,7 +106,7 @@
     )
 
 base_dest_cmd_fns: t.List[t.Callable[[], "BaseDestCmd"]] = [
-    astra_base_dest_cmd,
+    astradb_base_dest_cmd,
     azure_base_dest_cmd,
     box_base_dest_cmd,
     chroma_base_dest_cmd,

unstructured/ingest/cli/cmds/astra.py → unstructured/ingest/cli/cmds/astradb.py

@@ -4,11 +4,11 @@
 import click
 
 from unstructured.ingest.cli.interfaces import CliConfig, Dict
-from unstructured.ingest.connector.astra import AstraWriteConfig, SimpleAstraConfig
+from unstructured.ingest.connector.astradb import AstraDBWriteConfig, SimpleAstraDBConfig
 
 
 @dataclass
-class AstraCliConfig(SimpleAstraConfig, CliConfig):
+class AstraDBCliConfig(SimpleAstraDBConfig, CliConfig):
     @staticmethod
     def get_cli_options() -> t.List[click.Option]:
         options = [
@@ -48,7 +48,7 @@ def get_cli_options() -> t.List[click.Option]:
 
 
 @dataclass
-class AstraCliWriteConfig(AstraWriteConfig, CliConfig):
+class AstraDBCliWriteConfig(AstraDBWriteConfig, CliConfig):
     @staticmethod
     def get_cli_options() -> t.List[click.Option]:
         options = [
@@ -81,8 +81,8 @@ def get_base_src_cmd():
     from unstructured.ingest.cli.base.src import BaseSrcCmd
 
     cmd_cls = BaseSrcCmd(
-        cmd_name="astra",
-        cli_config=AstraCliConfig,
+        cmd_name="astradb",
+        cli_config=AstraDBCliConfig,
     )
     return cmd_cls
 
@@ -91,9 +91,9 @@ def get_base_dest_cmd():
     from unstructured.ingest.cli.base.dest import BaseDestCmd
 
     cmd_cls = BaseDestCmd(
-        cmd_name="astra",
-        cli_config=AstraCliConfig,
-        additional_cli_options=[AstraCliWriteConfig],
-        write_config=AstraWriteConfig,
+        cmd_name="astradb",
+        cli_config=AstraDBCliConfig,
+        additional_cli_options=[AstraDBCliWriteConfig],
+        write_config=AstraDBWriteConfig,
     )
     return cmd_cls

unstructured/ingest/connector/astra.py → unstructured/ingest/connector/astradb.py

@@ -31,23 +31,23 @@
 
 
 @dataclass
-class AstraAccessConfig(AccessConfig):
+class AstraDBAccessConfig(AccessConfig):
     token: str = enhanced_field(sensitive=True)
     api_endpoint: str = enhanced_field(sensitive=True)
 
 
 @dataclass
-class SimpleAstraConfig(BaseConnectorConfig):
-    access_config: AstraAccessConfig
+class SimpleAstraDBConfig(BaseConnectorConfig):
+    access_config: AstraDBAccessConfig
     collection_name: str
     namespace: t.Optional[str] = None
 
 
 @dataclass
-class AstraIngestDoc(IngestDocCleanupMixin, BaseSingleIngestDoc):
-    connector_config: SimpleAstraConfig
+class AstraDBIngestDoc(IngestDocCleanupMixin, BaseSingleIngestDoc):
+    connector_config: SimpleAstraDBConfig
     metadata: t.Dict[str, str] = field(default_factory=dict)
-    registry_name: str = "astra"
+    registry_name: str = "astradb"
 
     @property
     def filename(self):
@@ -76,7 +76,7 @@ def update_source_metadata(self, **kwargs):
         )
 
     @SourceConnectionError.wrap
-    @requires_dependencies(["astrapy"], extras="astra")
+    @requires_dependencies(["astrapy"], extras="astradb")
     @BaseSingleIngestDoc.skip_if_file_exists
     def get_file(self):
         self.filename.parent.mkdir(parents=True, exist_ok=True)
@@ -90,19 +90,19 @@ def get_file(self):
 
 
 @dataclass
-class AstraSourceConnector(SourceConnectorCleanupMixin, BaseSourceConnector):
-    connector_config: SimpleAstraConfig
+class AstraDBSourceConnector(SourceConnectorCleanupMixin, BaseSourceConnector):
+    connector_config: SimpleAstraDBConfig
     _astra_db: t.Optional["AstraDB"] = field(init=False, default=None)
     _astra_db_collection: t.Optional["AstraDBCollection"] = field(init=False, default=None)
 
     @property
-    @requires_dependencies(["astrapy"], extras="astra")
+    @requires_dependencies(["astrapy"], extras="astradb")
     def astra_db_collection(self) -> "AstraDBCollection":
         if self._astra_db_collection is None:
             from astrapy.db import AstraDB
 
             # Build the Astra DB object.
-            # caller_name/version for AstraDB tracking
+            # caller_name/version for Astra DB tracking
             self._astra_db = AstraDB(
                 api_endpoint=self.connector_config.access_config.api_endpoint,
                 token=self.connector_config.access_config.token,
@@ -117,27 +117,27 @@ def astra_db_collection(self) -> "AstraDBCollection":
             )
         return self._astra_db_collection  # type: ignore
 
-    @requires_dependencies(["astrapy"], extras="astra")
+    @requires_dependencies(["astrapy"], extras="astradb")
     @SourceConnectionError.wrap  # type: ignore
     def initialize(self):
         _ = self.astra_db_collection
 
-    @requires_dependencies(["astrapy"], extras="astra")
+    @requires_dependencies(["astrapy"], extras="astradb")
     def check_connection(self):
         try:
             _ = self.astra_db_collection
         except Exception as e:
             logger.error(f"Failed to validate connection {e}", exc_info=True)
             raise SourceConnectionError(f"failed to validate connection: {e}")
 
-    @requires_dependencies(["astrapy"], extras="astra")
+    @requires_dependencies(["astrapy"], extras="astradb")
     def get_ingest_docs(self):  # type: ignore
         # Perform the find operation
         astra_docs = list(self.astra_db_collection.paginated_find())
 
         doc_list = []
         for record in astra_docs:
-            doc = AstraIngestDoc(
+            doc = AstraDBIngestDoc(
                 connector_config=self.connector_config,
                 processor_config=self.processor_config,
                 read_config=self.read_config,
@@ -152,16 +152,16 @@ def get_ingest_docs(self):  # type: ignore
 
 
 @dataclass
-class AstraWriteConfig(WriteConfig):
+class AstraDBWriteConfig(WriteConfig):
     embedding_dimension: int
     requested_indexing_policy: t.Optional[t.Dict[str, t.Any]] = None
     batch_size: int = 20
 
 
 @dataclass
-class AstraDestinationConnector(BaseDestinationConnector):
-    write_config: AstraWriteConfig
-    connector_config: SimpleAstraConfig
+class AstraDBDestinationConnector(BaseDestinationConnector):
+    write_config: AstraDBWriteConfig
+    connector_config: SimpleAstraDBConfig
     _astra_db: t.Optional["AstraDB"] = field(init=False, default=None)
     _astra_db_collection: t.Optional["AstraDBCollection"] = field(init=False, default=None)
 
@@ -180,19 +180,19 @@ def to_dict(self, **kwargs):
         return _asdict(self_cp, **kwargs)
 
     @property
-    @requires_dependencies(["astrapy"], extras="astra")
+    @requires_dependencies(["astrapy"], extras="astradb")
     def astra_db_collection(self) -> "AstraDBCollection":
         if self._astra_db_collection is None:
             from astrapy.db import AstraDB
 
             collection_name = self.connector_config.collection_name
             embedding_dimension = self.write_config.embedding_dimension
 
-            # If the user has requested an indexing policy, pass it to the AstraDB
+            # If the user has requested an indexing policy, pass it to the Astra DB
             requested_indexing_policy = self.write_config.requested_indexing_policy
             options = {"indexing": requested_indexing_policy} if requested_indexing_policy else None
 
-            # caller_name/version for AstraDB tracking
+            # caller_name/version for Astra DB tracking
             self._astra_db = AstraDB(
                 api_endpoint=self.connector_config.access_config.api_endpoint,
                 token=self.connector_config.access_config.token,
@@ -209,12 +209,12 @@ def astra_db_collection(self) -> "AstraDBCollection":
             )
         return self._astra_db_collection
 
-    @requires_dependencies(["astrapy"], extras="astra")
+    @requires_dependencies(["astrapy"], extras="astradb")
     @DestinationConnectionError.wrap
     def initialize(self):
         _ = self.astra_db_collection
 
-    @requires_dependencies(["astrapy"], extras="astra")
+    @requires_dependencies(["astrapy"], extras="astradb")
     def check_connection(self):
         try:
             _ = self.astra_db_collection
@@ -223,7 +223,7 @@ def check_connection(self):
             raise DestinationConnectionError(f"failed to validate connection: {e}")
 
     def write_dict(self, *args, elements_dict: t.List[t.Dict[str, t.Any]], **kwargs) -> None:
-        logger.info(f"Inserting / updating {len(elements_dict)} documents to Astra.")
+        logger.info(f"Inserting / updating {len(elements_dict)} documents to Astra DB.")
 
         astra_batch_size = self.write_config.batch_size
 

unstructured/ingest/connector/registry.py

@@ -2,7 +2,7 @@
 from typing import Dict, Type, cast
 
 from unstructured.ingest.connector.airtable import AirtableIngestDoc
-from unstructured.ingest.connector.astra import AstraIngestDoc
+from unstructured.ingest.connector.astradb import AstraDBIngestDoc
 from unstructured.ingest.connector.biomed import BiomedIngestDoc
 from unstructured.ingest.connector.confluence import ConfluenceIngestDoc
 from unstructured.ingest.connector.delta_table import DeltaTableIngestDoc
@@ -46,7 +46,7 @@
 
 INGEST_DOC_NAME_TO_CLASS: Dict[str, Type[EnhancedDataClassJsonMixin]] = {
     "airtable": AirtableIngestDoc,
-    "astra": AstraIngestDoc,
+    "astradb": AstraDBIngestDoc,
     "azure": AzureBlobStorageIngestDoc,
     "biomed": BiomedIngestDoc,
     "box": BoxIngestDoc,