Add test.epub to docs

.gitignore

@@ -1,3 +1,6 @@
 .venv
 db
 dummy
+.ENV
+.env
+__pycache__

app.py

@@ -289,12 +289,12 @@ def ingest(
     ]
 
 
+# TheBloke/Wizard-Vicuna-7B-Uncensored-HF
 # https://huggingface.co/TheBloke/vicuna-7B-1.1-HF
 def gen_local_llm(model_id="TheBloke/vicuna-7B-1.1-HF"):
     """Gen a local llm.
 
     localgpt run_localgpt
-
     https://medium.com/pytorch/bettertransformer-out-of-the-box-performance-for-huggingface-transformers-3fbe27d50ab2
     with torch.device(“cuda”):
         model = AutoModelForCausalLM.from_pretrained(“gpt2-large”, torch_dtype=torch.float16)
@@ -354,7 +354,9 @@ def load_qa(device=None, model_name: str = "hkunlp/instructor-base"):
     llm = gen_local_llm()  # "TheBloke/vicuna-7B-1.1-HF" 12G?
 
     qa = RetrievalQA.from_chain_type(
-        llm=llm, chain_type="stuff", retriever=retriever, return_source_documents=True
+        llm=llm, chain_type="stuff", 
+        retriever=retriever, 
+        return_source_documents=True,
     )
 
     logger.info("Done qa")

epub_loader.py

@@ -0,0 +1,38 @@
+"""Loads an epub file into a list of documents."""
+from dataclasses import dataclass
+from pathlib import Path
+from typing import List, Union
+
+from epub2txt import epub2txt
+from langchain.docstore.document import Document
+from langchain.document_loaders.base import BaseLoader
+from loguru import logger
+
+
+@dataclass
+class EpubLoader(BaseLoader):
+    """Load an epub file into a list of documents.
+
+    Args:
+        file_path: file path or url to epub
+    Returns:
+        self.load() -> list of Documents
+    """
+    file_path: Union[str, Path]
+
+    def load(self) -> List[Document]:
+        """Load data into document objects."""
+        try:
+            texts = epub2txt(self.file_path, outputlist=True)
+            ch_titles = epub2txt.content_titles
+
+        except Exception as exc:
+            logger.error(exc)
+            raise
+
+        docs = []
+        for title, text in zip(ch_titles, texts):
+            metadata = {"source": self.file_path, "ch.": title}
+            docs.append(Document(page_content=text, metadata=metadata))
+
+        return docs