Getting a doc_id and filename map

[shneeba]

Before I go submitting a pull request I just want to see if there's already a way to get a doc_id and filename mapping?

I added the following to the Query class:

    def get_docid_filename_map(self):
        """
        Returns a list of dictionaries, each containing "doc_ID" and "file_source"
        for every distinct doc_ID in this library.

        The mapping is one doc_ID -> one file_source (the first occurrence in
        the collection). This is usually enough to uniquely match a doc_ID to
        a particular file.

        Returns
        -------
        list of dict
            Each dict has the form:
                {
                "doc_ID": <int or str>,
                "file_source": <str>
                }
            where <int or str> is the document ID, and <str> is the associated filename or path.
        """
        # 1) Create a CollectionRetrieval object for this library
        collection_retrieval = CollectionRetrieval(self.library_name, account_name=self.account_name)
        
        # 2) Retrieve the entire collection cursor
        collection_cursor = collection_retrieval.get_whole_collection()
        
        # 3) Pull all results from the cursor
        all_entries = collection_cursor.pull_all()

        # 4) Build a dictionary to store doc_ID -> file_source
        doc_map = {}

        for entry in all_entries:
            doc_id = entry.get("doc_ID")
            file_source = entry.get("file_source")

            # Only set file_source once per doc_ID (first occurrence)
            if doc_id is not None and file_source is not None:
                if doc_id not in doc_map:
                    doc_map[doc_id] = file_source

        # 5) Convert the mapping into a list of dicts
        result = [
            {"doc_ID": did, "file_source": fname}
            for did, fname in doc_map.items()
        ]

        # 6) Close underlying database cursor/connection
        collection_retrieval.close()

        # 7) Return the final list
        return result

This seems to work quite well for what I'm doing. If there's a better way let me know, if not I'll submit the pull request!

Cheers