forked from redis-developer/ArXivChatGuru
-
Notifications
You must be signed in to change notification settings - Fork 0
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Langchain integration (redis-developer#3)
* use langchain * finish integration
- Loading branch information
1 parent
b747e11
commit cdcb494
Showing
8 changed files
with
158 additions
and
344 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -1,99 +1 @@ | ||
|
|
||
| import typing as t | ||
|
|
||
| from . import db | ||
| from .models import ( | ||
| MAX_SECTION_LEN, | ||
| SEPARATOR, | ||
| SEPARATOR_LEN, | ||
| get_embedding, | ||
| get_completion | ||
| ) | ||
|
|
||
| redis_conn = db.init() | ||
| PROMPT_HEADER = """Answer the question as truthfully as possible using the provided context, and if the answer is not contained within the text below, say "I don't know."\n\nContext:\n""" | ||
|
|
||
| def search_semantic_redis(search_query: str, n: int) -> t.List[dict]: | ||
| """ | ||
| Search Redis using computed embeddings from OpenAI. | ||
| Args: | ||
| search_query (str): Text query to embed and use in document retrieval. | ||
| n (int): Number of documents to consider. | ||
| Returns: | ||
| list<dict>: List of relevant documents ordered by similarity score. | ||
| """ | ||
| embedding = get_embedding(text=search_query) | ||
| return db.search_redis( | ||
| redis_conn, | ||
| query_vector=embedding, | ||
| k=n, | ||
| return_fields=["title", "content", "tokens"] | ||
| ) | ||
|
|
||
| def construct_prompt(question: str) -> str: | ||
| """ | ||
| Construct full prompt based on the input question using | ||
| the document sections indexed in Redis. | ||
| Args: | ||
| question (str): User input question. | ||
| pre_filter (str, optional): Pre filter to constrain the KNN search with conditions. | ||
| Returns: | ||
| str: Full prompt string to pass along to a generative language model. | ||
| """ | ||
| chosen_sections = [] | ||
| chosen_sections_len = 0 | ||
| chosen_sections_indexes = [] | ||
|
|
||
| # Search for relevant document sections based on the question | ||
| most_relevant_document_sections = search_semantic_redis(question, n = 5) | ||
|
|
||
| # Iterate through results | ||
| for document_section in most_relevant_document_sections: | ||
| # Add contexts until we run out of token space | ||
| chosen_sections_len += int(document_section['tokens']) + SEPARATOR_LEN | ||
| if chosen_sections_len > MAX_SECTION_LEN: | ||
| break | ||
|
|
||
| chosen_sections.append(SEPARATOR + document_section['content'].replace("\n", " ")) | ||
| chosen_sections_indexes.append(document_section['id']) | ||
|
|
||
| # Useful diagnostic information | ||
| print(f"Selected {len(chosen_sections)} document sections:") | ||
| print("\n".join(chosen_sections_indexes)) | ||
|
|
||
| return PROMPT_HEADER + "".join(chosen_sections) + "\n\n Q: " + question + "\n A:" | ||
|
|
||
| def answer_question_with_context( | ||
| question: str, | ||
| show_prompt: bool = False, | ||
| explicit_prompt: str = "", | ||
| tokens_response=100, | ||
| temperature=0.0 | ||
| ) -> str: | ||
| """ | ||
| Answer the question. | ||
| Args: | ||
| question (str): Input question from the user. | ||
| show_prompt (bool, optional): Print out the prompt? Defaults to False. | ||
| explicit_prompt (str, optional): Use an explicit prompt provided by user? Defaults to "". | ||
| tokens_response (int, optional): Max number of tokens in the response. Defaults to 100. | ||
| temperature (float, optional): Model temperature. Defaults to 0.0. | ||
| Returns: | ||
| str: _description_ | ||
| """ | ||
| if explicit_prompt == "": | ||
| # Construct prompt with Redis Vector Search | ||
| prompt = construct_prompt(question) | ||
| else: | ||
| prompt = f"{explicit_prompt}\n\n{question}" | ||
|
|
||
| if show_prompt: | ||
| print(prompt) | ||
|
|
||
| return get_completion(prompt, max_tokens=tokens_response, temperature=temperature) | ||
| from .llm import make_qna_chain |
This file was deleted.
Oops, something went wrong.
Oops, something went wrong.