unstructured-ingest 0.6.4__py3-none-any.whl → 0.7.1__py3-none-any.whl

unstructured_ingest/connector/biomed.py

@@ -1,320 +0,0 @@
-import os
-import typing as t
-import urllib.request
-from dataclasses import dataclass
-from ftplib import FTP, error_perm
-from pathlib import Path
-
-from unstructured_ingest.error import SourceConnectionError, SourceConnectionNetworkError
-from unstructured_ingest.interfaces import (
-    BaseConnectorConfig,
-    BaseSingleIngestDoc,
-    BaseSourceConnector,
-    IngestDocCleanupMixin,
-    SourceConnectorCleanupMixin,
-)
-from unstructured_ingest.logger import logger
-from unstructured_ingest.utils.data_prep import (
-    validate_date_args,
-)
-from unstructured_ingest.utils.dep_check import requires_dependencies
-
-if t.TYPE_CHECKING:
-    from requests import Response, Session
-
-DOMAIN = "ftp.ncbi.nlm.nih.gov"
-FTP_DOMAIN = f"ftp://{DOMAIN}"
-PMC_DIR = "pub/pmc"
-PDF_DIR = "oa_pdf"
-
-
-@dataclass
-class BiomedFileMeta:
-    ftp_path: str
-    download_filepath: str
-    output_filepath: str
-
-
-@dataclass
-class SimpleBiomedConfig(BaseConnectorConfig):
-    """Connector config where path is the FTP directory path and
-    id_, from_, until, format are API parameters."""
-
-    path: t.Optional[str] = None
-    # OA Web Service API Options
-    api_id: t.Optional[str] = None
-    api_from: t.Optional[str] = None
-    api_until: t.Optional[str] = None
-    max_request_time: int = 45
-
-    def validate_api_inputs(self):
-        valid = False
-
-        if self.api_from:
-            valid = validate_date_args(self.api_from)
-
-        if self.api_until:
-            valid = validate_date_args(self.api_until)
-
-        return valid
-
-    def __post_init__(self):
-        self.is_file = False
-        self.is_dir = False
-        self.is_api = False
-
-        if not self.path:
-            is_valid = self.validate_api_inputs()
-            if not is_valid:
-                raise ValueError(
-                    "Path argument or at least one of the "
-                    "OA Web Service arguments MUST be provided.",
-                )
-
-            self.is_api = True
-        else:
-            self.path = self.path.strip("/")
-            is_valid = self.path.lower().startswith(PDF_DIR)
-
-            if not is_valid:
-                raise ValueError(f"Path MUST start with {PDF_DIR}")
-
-            ftp = FTP(DOMAIN)
-            ftp.login()
-
-            path = Path(PMC_DIR) / self.path
-            response = ""
-            try:
-                if path.suffix == ".pdf":
-                    response = ftp.cwd(str(path.parent))
-                    self.is_file = True
-                else:
-                    response = ftp.cwd(str(path))
-            except error_perm as exc:
-                if "no such file or directory" in exc.args[0].lower():
-                    raise ValueError(f"The path: {path} is not valid.")
-                elif "not a directory" in exc.args[0].lower():
-                    self.is_file = True
-                elif "command successful" in response:
-                    self.is_dir = True
-                else:
-                    raise ValueError(
-                        "Something went wrong when validating the path: {path}.",
-                    )
-
-
-@dataclass
-class BiomedIngestDoc(IngestDocCleanupMixin, BaseSingleIngestDoc):
-    connector_config: SimpleBiomedConfig
-    file_meta: BiomedFileMeta
-    registry_name: str = "biomed"
-
-    @property
-    def filename(self):
-        return Path(self.file_meta.download_filepath).resolve()  # type: ignore
-
-    @property
-    def _output_filename(self):
-        return Path(f"{self.file_meta.output_filepath}.json").resolve()
-
-    def cleanup_file(self):
-        if (
-            not self.read_config.preserve_downloads
-            and self.filename.is_file()
-            and not self.read_config.download_only
-        ):
-            logger.debug(f"cleaning up {self}")
-            Path.unlink(self.filename)
-
-    @SourceConnectionError.wrap
-    @BaseSingleIngestDoc.skip_if_file_exists
-    def get_file(self):
-        download_path = self.file_meta.download_filepath  # type: ignore
-        dir_ = Path(os.path.dirname(download_path))  # type: ignore
-        if not dir_.is_dir():
-            logger.debug(f"creating directory: {dir_}")
-
-            if dir_:
-                dir_.mkdir(parents=True, exist_ok=True)
-        self._retrieve()
-        logger.debug(f"file downloaded: {self.file_meta.download_filepath}")
-
-    @SourceConnectionNetworkError.wrap
-    def _retrieve(self):
-        urllib.request.urlretrieve(
-            self.file_meta.ftp_path,  # type: ignore
-            self.file_meta.download_filepath,
-        )
-
-
-class BiomedSourceConnector(SourceConnectorCleanupMixin, BaseSourceConnector):
-    """Objects of this class support fetching documents from Biomedical literature FTP directory"""
-
-    connector_config: SimpleBiomedConfig
-
-    def get_base_endpoints_url(self) -> str:
-        endpoint_url = "https://www.ncbi.nlm.nih.gov/pmc/utils/oa/oa.fcgi?format=pdf"
-
-        if self.connector_config.api_id:
-            endpoint_url += f"&id={self.connector_config.api_id}"
-
-        if self.connector_config.api_from:
-            endpoint_url += f"&from={self.connector_config.api_from}"
-
-        if self.connector_config.api_until:
-            endpoint_url += f"&until={self.connector_config.api_until}"
-
-        return endpoint_url
-
-    @requires_dependencies(["requests"], extras="biomed")
-    def _list_objects_api(self) -> t.List[BiomedFileMeta]:
-        from bs4 import BeautifulSoup
-        from requests import Session
-        from requests.adapters import HTTPAdapter
-
-        def urls_to_metadata(urls):
-            files = []
-            for url in urls:
-                parts = url.split(PDF_DIR)
-                if len(parts) > 1:
-                    local_path = parts[1].strip("/")
-                    files.append(
-                        BiomedFileMeta(
-                            ftp_path=url,
-                            download_filepath=(Path(self.read_config.download_dir) / local_path)
-                            .resolve()
-                            .as_posix(),
-                            output_filepath=(Path(self.processor_config.output_dir) / local_path)
-                            .resolve()
-                            .as_posix(),
-                        ),
-                    )
-
-            return files
-
-        files: t.List[BiomedFileMeta] = []
-
-        endpoint_url = self.get_base_endpoints_url()
-
-        while endpoint_url:
-            session = Session()
-            adapter = HTTPAdapter()
-            session.mount("http://", adapter)
-            session.mount("https://", adapter)
-            response = self._get_request(session=session, endpoint_url=endpoint_url)
-            soup = BeautifulSoup(response.content, features="lxml")
-            urls = [link["href"] for link in soup.find_all("link")]
-
-            if not urls:
-                return files
-
-            endpoint_url = urls[-1] if "resumptiontoken" in urls[-1].lower() else None
-            if endpoint_url:
-                urls = urls[:-1]
-
-            files.extend(urls_to_metadata(urls))
-
-        return files
-
-    @SourceConnectionNetworkError.wrap
-    def _get_request(self, session: "Session", endpoint_url: str) -> "Response":
-        return session.get(endpoint_url, timeout=self.connector_config.max_request_time)
-
-    def _list_objects(self) -> t.List[BiomedFileMeta]:
-        files = []
-
-        # Conform to mypy, null check performed elsewhere.
-        # Wouldn't be in this method unless self.config.path exists
-        path: str = self.connector_config.path if self.connector_config.path else ""
-
-        def traverse(path, download_dir, output_dir):
-            full_path = Path(PMC_DIR) / path
-            logger.debug(f"traversing directory: {full_path}")
-
-            ftp = FTP(DOMAIN)
-            ftp.login()
-
-            try:
-                response = ftp.cwd(str(full_path))
-            except error_perm:
-                raise ValueError(f"{full_path} is not a valid directory.")
-
-            if "command successful" in response.lower():
-                sub_paths = [path / p for p in ftp.nlst()]
-
-                if not sub_paths:
-                    return
-
-                ext = Path(sub_paths[0]).suffix
-                if ext:
-                    for sub_path in sub_paths:
-                        ftp_path = f"{FTP_DOMAIN}/{PMC_DIR}/{sub_path}"
-                        local_path = "/".join(str(sub_path).split("/")[1:])
-                        files.append(
-                            BiomedFileMeta(
-                                ftp_path=ftp_path,
-                                download_filepath=(Path(self.read_config.download_dir) / local_path)
-                                .resolve()
-                                .as_posix(),
-                                output_filepath=(
-                                    Path(self.processor_config.output_dir) / local_path
-                                )
-                                .resolve()
-                                .as_posix(),
-                            ),
-                        )
-
-                else:
-                    for sub_path in sub_paths:
-                        traverse(sub_path, download_dir, output_dir)
-
-            else:
-                raise ValueError(f"{full_path} is not a valid directory.")
-
-        ftp_path = f"{FTP_DOMAIN}/{PMC_DIR}/{self.connector_config.path}"
-        if self.connector_config.is_file:
-            local_path = "/".join(path.split("/")[1:])
-            return [
-                BiomedFileMeta(
-                    ftp_path=ftp_path,
-                    download_filepath=(Path(self.read_config.download_dir) / local_path)
-                    .resolve()
-                    .as_posix(),
-                    output_filepath=(Path(self.processor_config.output_dir) / local_path)
-                    .resolve()
-                    .as_posix(),
-                ),
-            ]
-        else:
-            traverse(
-                Path(path),
-                Path(self.read_config.download_dir),
-                Path(self.processor_config.output_dir),
-            )
-
-        return files
-
-    def initialize(self):
-        pass
-
-    @requires_dependencies(["requests"], extras="biomed")
-    def check_connection(self):
-        import requests
-
-        resp = requests.head(self.get_base_endpoints_url())
-        try:
-            resp.raise_for_status()
-        except requests.HTTPError as http_error:
-            raise SourceConnectionError(f"failed to validate connection: {http_error}")
-
-    def get_ingest_docs(self):
-        files = self._list_objects_api() if self.connector_config.is_api else self._list_objects()
-        return [
-            BiomedIngestDoc(
-                processor_config=self.processor_config,
-                connector_config=self.connector_config,
-                read_config=self.read_config,
-                file_meta=file,
-            )
-            for file in files
-        ]

unstructured_ingest/connector/chroma.py

@@ -1,158 +0,0 @@
-import copy
-import typing as t
-import uuid
-from dataclasses import dataclass
-
-from unstructured_ingest.enhanced_dataclass.core import _asdict
-from unstructured_ingest.error import DestinationConnectionError
-from unstructured_ingest.interfaces import (
-    AccessConfig,
-    BaseConnectorConfig,
-    BaseDestinationConnector,
-    WriteConfig,
-)
-from unstructured_ingest.logger import logger
-from unstructured_ingest.utils.data_prep import batch_generator, flatten_dict
-from unstructured_ingest.utils.dep_check import requires_dependencies
-
-if t.TYPE_CHECKING:
-    from chromadb import Collection as ChromaCollection
-
-
-@dataclass
-class ChromaAccessConfig(AccessConfig):
-    settings: t.Optional[t.Dict[str, str]] = None
-    headers: t.Optional[t.Dict[str, str]] = None
-
-
-@dataclass
-class SimpleChromaConfig(BaseConnectorConfig):
-    access_config: ChromaAccessConfig
-    collection_name: str
-    path: t.Optional[str] = None
-    tenant: t.Optional[str] = "default_tenant"
-    database: t.Optional[str] = "default_database"
-    host: t.Optional[str] = None
-    port: t.Optional[int] = None
-    ssl: bool = False
-
-
-@dataclass
-class ChromaWriteConfig(WriteConfig):
-    batch_size: int = 100
-
-
-@dataclass
-class ChromaDestinationConnector(BaseDestinationConnector):
-    write_config: ChromaWriteConfig
-    connector_config: SimpleChromaConfig
-    _collection: t.Optional["ChromaCollection"] = None
-
-    @property
-    def chroma_collection(self):
-        if self._collection is None:
-            self._collection = self.create_collection()
-        return self._collection
-
-    def initialize(self):
-        pass
-
-    @DestinationConnectionError.wrap
-    def check_connection(self):
-        _ = self.chroma_collection
-
-    def to_dict(self, **kwargs):
-        """
-        The _collection variable in this dataclass breaks deepcopy due to:
-        TypeError: cannot pickle 'module' object
-        When serializing, remove it, meaning collection data will need to be reinitialized
-        when deserialized
-        """
-        self_cp = copy.copy(self)
-        if hasattr(self_cp, "_collection"):
-            setattr(self_cp, "_collection", None)
-        return _asdict(self_cp, **kwargs)
-
-    @requires_dependencies(["chromadb"], extras="chroma")
-    def create_collection(self) -> "ChromaCollection":
-        import chromadb
-
-        if self.connector_config.path:
-            chroma_client = chromadb.PersistentClient(
-                path=self.connector_config.path,
-                settings=self.connector_config.settings,
-                tenant=self.connector_config.tenant,
-                database=self.connector_config.database,
-            )
-
-        elif self.connector_config.host and self.connector_config.port:
-            chroma_client = chromadb.HttpClient(
-                host=self.connector_config.host,
-                port=self.connector_config.port,
-                ssl=self.connector_config.ssl,
-                headers=self.connector_config.access_config.headers,
-                settings=self.connector_config.access_config.settings,
-                tenant=self.connector_config.tenant,
-                database=self.connector_config.database,
-            )
-        else:
-            raise ValueError("Chroma connector requires either path or host and port to be set.")
-
-        collection = chroma_client.get_or_create_collection(
-            name=self.connector_config.collection_name
-        )
-        return collection
-
-    @DestinationConnectionError.wrap
-    @requires_dependencies(["chromadb"], extras="chroma")
-    def upsert_batch(self, batch):
-        collection = self.chroma_collection
-
-        try:
-            # Chroma wants lists even if there is only one element
-            # Upserting to prevent duplicates
-            collection.upsert(
-                ids=batch["ids"],
-                documents=batch["documents"],
-                embeddings=batch["embeddings"],
-                metadatas=batch["metadatas"],
-            )
-        except Exception as e:
-            raise ValueError(f"chroma error: {e}") from e
-
-    @staticmethod
-    def prepare_chroma_list(chunk: t.Tuple[t.Dict[str, t.Any]]) -> t.Dict[str, t.List[t.Any]]:
-        """Helper function to break a tuple of dicts into list of parallel lists for ChromaDb.
-        ({'id':1}, {'id':2}, {'id':3}) -> {'ids':[1,2,3]}"""
-        chroma_dict = {}
-        chroma_dict["ids"] = [x.get("id") for x in chunk]
-        chroma_dict["documents"] = [x.get("document") for x in chunk]
-        chroma_dict["embeddings"] = [x.get("embedding") for x in chunk]
-        chroma_dict["metadatas"] = [x.get("metadata") for x in chunk]
-        # Make sure all lists are of the same length
-        assert (
-            len(chroma_dict["ids"])
-            == len(chroma_dict["documents"])
-            == len(chroma_dict["embeddings"])
-            == len(chroma_dict["metadatas"])
-        )
-        return chroma_dict
-
-    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 destination ")
-
-        chroma_batch_size = self.write_config.batch_size
-
-        for chunk in batch_generator(elements_dict, chroma_batch_size):
-            self.upsert_batch(self.prepare_chroma_list(chunk))
-
-    def normalize_dict(self, element_dict: dict) -> dict:
-        element_id = element_dict.get("element_id", str(uuid.uuid4()))
-        return {
-            "id": element_id,
-            "embedding": element_dict.pop("embeddings", None),
-            "document": element_dict.pop("text", None),
-            "metadata": flatten_dict(
-                element_dict, separator="-", flatten_lists=True, remove_none=True
-            ),
-        }

unstructured_ingest/connector/clarifai.py

@@ -1,122 +0,0 @@
-import typing as t
-import uuid
-from dataclasses import dataclass, field
-
-from unstructured_ingest.enhanced_dataclass import enhanced_field
-from unstructured_ingest.error import DestinationConnectionError
-from unstructured_ingest.interfaces import (
-    AccessConfig,
-    BaseConnectorConfig,
-    BaseDestinationConnector,
-    WriteConfig,
-)
-from unstructured_ingest.logger import logger
-from unstructured_ingest.utils.data_prep import flatten_dict
-from unstructured_ingest.utils.dep_check import requires_dependencies
-
-if t.TYPE_CHECKING:
-    from clarifai.client.input import Inputs
-
-
-@dataclass
-class ClarifaiAccessConfig(AccessConfig):
-    api_key: str = enhanced_field(sensitive=True)
-
-
-@dataclass
-class SimpleClarifaiConfig(BaseConnectorConfig):
-    access_config: ClarifaiAccessConfig
-    app_id: str
-    user_id: str
-    dataset_id: t.Optional[str] = None
-
-
-@dataclass
-class ClarifaiWriteConfig(WriteConfig):
-    batch_size: int = 50
-
-
-@dataclass
-class ClarifaiDestinationConnector(BaseDestinationConnector):
-    write_config: ClarifaiWriteConfig
-    connector_config: SimpleClarifaiConfig
-    _client: t.Optional["Inputs"] = field(init=False, default=None)
-
-    @property
-    @requires_dependencies(["clarifai"], extras="clarifai")
-    def client(self) -> "Inputs":
-        if self._client is None:
-            from clarifai.client.input import Inputs
-
-            access_conf = self.connector_config.access_config
-            try:
-                if access_conf.api_key is not None:
-                    clarifai_pat = access_conf.api_key
-            except Exception as e:
-                raise (f"please provide clarifai PAT key : {e}")
-
-            self._client = Inputs(
-                app_id=self.connector_config.app_id,
-                user_id=self.connector_config.user_id,
-                pat=clarifai_pat,
-            )
-        return self._client
-
-    @requires_dependencies(["clarifai"], extras="clarifai")
-    @DestinationConnectionError.wrap
-    def initialize(self):
-        _ = self.client
-
-    def check_connection(self):
-        try:
-            _ = [inp for inp in self.client.list_inputs(page_no=1, per_page=1)]  # noqa: C416
-        except Exception as e:
-            logger.error(f"Failed to validate connection {e}", exc_info=True)
-            raise DestinationConnectionError(f"failed to validate connection: {e}")
-
-    def normalize_dict(self, element_dict: dict) -> dict:
-        """Modifying schema of the dict in order to compile with clarifai input formats"""
-        return {
-            "input_id": str(uuid.uuid4().hex),
-            "text": element_dict.pop("text", None),
-            "metadata": {
-                **flatten_dict(
-                    element_dict,
-                    separator="_",
-                    flatten_lists=True,
-                    remove_none=True,
-                ),
-            },
-        }
-
-    def write_dict(self, *args, elements_dict: t.List[t.Dict[str, t.Any]], **kwargs) -> None:
-        from google.protobuf.struct_pb2 import Struct
-
-        logger.info(
-            f"writing {len(elements_dict)} objects to destination "
-            f"app {self.connector_config.app_id} "
-        )
-        try:
-            batch_size = self.write_config.batch_size
-            for idx in range(0, len(elements_dict), batch_size):
-                batch_dict = elements_dict[idx : batch_size + idx]
-                input_batch = []
-                for elem in batch_dict:
-                    meta_struct = Struct()
-                    meta_struct.update(elem["metadata"])
-                    input_batch.append(
-                        self._client.get_text_input(
-                            input_id=elem["input_id"],
-                            raw_text=elem["text"],
-                            dataset_id=self.connector_config.dataset_id,
-                            metadata=meta_struct,
-                        )
-                    )
-                result_id = self._client.upload_inputs(inputs=input_batch)
-                logger.debug(
-                    f"Input posted successfully into {self.connector_config.app_id}. \
-                    Result id: {result_id}"
-                )
-
-        except Exception as e:
-            raise e