|
| 1 | +```python Python Ingest v2 |
| 2 | +import os |
| 3 | + |
| 4 | +from unstructured_ingest.v2.pipeline.pipeline import Pipeline |
| 5 | +from unstructured_ingest.v2.interfaces import ProcessorConfig |
| 6 | + |
| 7 | +from unstructured_ingest.v2.processes.connectors.outlook import ( |
| 8 | + OutlookIndexerConfig, |
| 9 | + OutlookDownloaderConfig, |
| 10 | + OutlookConnectionConfig, |
| 11 | + OutlookAccessConfig |
| 12 | +) |
| 13 | + |
| 14 | +from unstructured_ingest.v2.processes.connectors.local import LocalUploaderConfig |
| 15 | +from unstructured_ingest.v2.processes.partitioner import PartitionerConfig |
| 16 | +from unstructured_ingest.v2.processes.chunker import ChunkerConfig |
| 17 | +from unstructured_ingest.v2.processes.embedder import EmbedderConfig |
| 18 | + |
| 19 | +# Embedding is optional. |
| 20 | + |
| 21 | +if __name__ == "__main__": |
| 22 | + Pipeline.from_configs( |
| 23 | + context=ProcessorConfig(), |
| 24 | + indexer_config=OutlookIndexerConfig( |
| 25 | + outlook_folders=["Inbox"], |
| 26 | + recursive=False, |
| 27 | + user_email=os.getenv("OUTLOOK_USER_EMAIL") |
| 28 | + ), |
| 29 | + downloader_config=OutlookDownloaderConfig(download_dir=os.getenv("LOCAL_FILE_DOWNLOAD_DIR")), |
| 30 | + source_connection_config=OutlookConnectionConfig( |
| 31 | + access_config=OutlookAccessConfig(client_cred=os.getenv("OUTLOOK_APP_CLIENT_SECRET")), |
| 32 | + client_id=os.getenv("OUTLOOK_APP_CLIENT_ID"), |
| 33 | + tenant=os.getenv("OUTLOOK_APP_TENANT") |
| 34 | + ), |
| 35 | + partitioner_config=PartitionerConfig( |
| 36 | + partition_by_api=True, |
| 37 | + api_key=os.getenv("UNSTRUCTURED_API_KEY"), |
| 38 | + partition_endpoint=os.getenv("UNSTRUCTURED_API_URL") |
| 39 | + ), |
| 40 | + chunker_config=ChunkerConfig(chunking_strategy="by_title"), |
| 41 | + embedder_config=EmbedderConfig(embedding_provider="huggingface"), |
| 42 | + uploader_config=LocalUploaderConfig(output_dir=os.getenv("LOCAL_FILE_OUTPUT_DIR")) |
| 43 | + ).run() |
| 44 | +``` |
0 commit comments