janus.utils.pdf_docs_reader#

Attributes#

log

Classes#

PDFDocsReader

Module Contents#

janus.utils.pdf_docs_reader.log#
class janus.utils.pdf_docs_reader.PDFDocsReader(language, chunk_size=1000, chunk_overlap=100, start_page=None, end_page=None, vectorizer=TfidfVectorizer())#
Parameters:

  • language (str) –

  • chunk_size (int) –

  • chunk_overlap (int) –

  • start_page (Optional[int]) –

  • end_page (Optional[int]) –

  • vectorizer (sklearn.feature_extraction.text.CountVectorizer) –

load_and_chunk_pdf()#
Return type:

List[str]

vectorize_documents()#
Return type:

(sklearn.feature_extraction.text.TfidfVectorizer, any)

search_language_reference(query, top_k=1, min_similarity=0.1)#

Searches through the vectorized PDF for the query using tf-idf and returns a list of langchain Documents.

Parameters:

  • query (List[str]) –

  • top_k (int) –

  • min_similarity (float) –

Return type:

List[langchain_core.documents.Document]