fix embeddings module

server/nlp/embeddings.py

@@ -1,3 +1,8 @@
+"""Embeddings.
+
+This module provides functions for embedding documents and querying them in Redis.
+"""
+
 import os
 import time
 
@@ -31,17 +36,13 @@
 
 
 def load_corpus(corpus: list[RedisDocument]):
-    """Loads given corpus into redis
+    """Loads given corpus into redis.
 
-    PARAMETERS
-    ----------
-    corpus : :obj:`list` of :obj:`RedisDocument`
-        list of documents, each represented by dictionary
+    Args:
+        corpus: list of documents, each represented by dictionary
 
     Raises:
-    ------
-    Exception
-        if failed to load corpus into redis
+        exception: if failed to load corpus into redis
     """
     print("loading corpus...")
 
@@ -57,6 +58,14 @@ def load_corpus(corpus: list[RedisDocument]):
 
 
 def compute_openai_embeddings(texts):
+    """Compute embeddings from texts using OpenAI.
+
+    Args:
+        texts: list of texts to embed
+
+    Returns:
+        list of embeddings
+    """
     embeddings = []
     for i in range(len(texts)):
         embeddings.append(
@@ -68,7 +77,7 @@ def compute_openai_embeddings(texts):
 
 
 def compute_embeddings():
-    """Compute embeddings from redis documents"""
+    """Compute embeddings from redis documents."""
     print("computing embeddings...")
 
     # get keys, questions, content
@@ -97,17 +106,14 @@ def compute_embeddings():
 
 
 def load_embeddings(embeddings: list[list[float]]):
-    """Load embeddings into redis
+    """Load embeddings into redis.
 
-    PARAMETERS
-    ----------
-    embeddings : :obj:`list` of :obj:`list` of :obj:`float`
-        list of embeddings
+    Args:
+        embeddings:
+            list of embeddings
 
     Raises:
-    ------
-    Exception
-        if failed to load embeddings into redis
+        exception: if failed to load embeddings into redis
     """
     print("loading embeddings into redis...")
 
@@ -125,18 +131,16 @@ def load_embeddings(embeddings: list[list[float]]):
 
 
 def create_index(corpus_len: int):
-    """Create search index in redis
-    assumes that documents and embeddings have already been loaded into redis
+    """Create search index in redis.
+
+    Assumes that documents and embeddings have already been loaded into redis
 
-    PARAMETERS
-    ----------
-    corpus_len : :obj:`int`
-        number of documents in corpus
+    Args:
+        corpus_len:
+            number of documents in corpus
 
     Raises:
-    ------
-    Exception
-        if failed to create index
+        exception: if failed to create index
     """
     print("creating index...")
 
@@ -178,12 +182,13 @@ def create_index(corpus_len: int):
 
 
 def create_query(k: int):
-    """Create k-NN redis query
+    """Create k-NN redis query.
 
-    PARAMETERS
-    ----------
-    k : :obj:`int`
-        number of nearest neighbors to return
+    Args:
+        k: number of nearest neighbors to return
+
+    Returns:
+        redis query object
     """
     return (
         Query(f"(*)=>[KNN {k} @vector $query_vector AS vector_score]")
@@ -194,18 +199,13 @@ def create_query(k: int):
 
 
 def queries(query, queries: list[str]) -> list[dict]:
-    """Run queries against redis
+    """Run queries against redis.
 
-    PARAMETERS
-    ----------
-    query : :obj:`Query`
-        redis query object
-    queries : :obj:`list` of :obj:`str`
-        list of question queries
+    Args:
+        query: redis query object
+        queries: list of question queries
 
     Returns:
-    -------
-    :obj:`list` of :obj:`dict`
         list of dictionaries containing query and result
     """
     print("running queries...")
@@ -243,36 +243,27 @@ def queries(query, queries: list[str]) -> list[dict]:
 
 
 def query_all(k: int, questions: list[str]):
-    """Return k most similar documents for each query
+    """Return k most similar documents for each query.
 
-    PARAMETERS
-    ----------
-    k : :obj:`int`
-        number of nearest neighbors to return
-    questions : :obj:`list` of :obj:`str`
-        list of question queries
+    Args:
+        k: number of nearest neighbors to return
+        questions: list of question queries
 
     Returns:
-    -------
-    :obj:`list` of :obj:`dict`
         list of dictionaries containing query and result
     """
     redis_query = create_query(k)
     return queries(redis_query, questions)
 
 
 def embed_corpus(corpus: list[RedisDocument]):
-    """Load corpus, compute embeddings, load embeddings into redis
+    """Load corpus, compute embeddings, load embeddings into redis.
 
-    PARAMETERS
-    ----------
-    corpus : :obj:`list` of :obj:`dict`
-        list of documents, each represented by dictionary
+    Args:
+        corpus: list of documents, each represented by dictionary
 
     Raises:
-    ------
-    Exception
-        if failed to load corpus
+        exception: if failed to load corpus
     """
     # flush database
     print("cleaning database...")
@@ -288,37 +279,38 @@ def embed_corpus(corpus: list[RedisDocument]):
     create_index(len(corpus))
 
 
-def test():
-    try:
-        embed_corpus()
-    except Exception as err:
-        print(f"Unexpected {err=}, {type(err)=}")
-        raise
-
-    questions = [
-        "What is the deadline to apply for the hackathon?",
-        "When is HackMIT?",
-        "What are the challenges?",
-        "How does judging work?",
-        "What building should I go to during the event?",
-        "What prizes are available?",
-        "How many people are allowed on a team?",
-        "What is HackMIT?",
-        "Can I attend HackMIT if I am an MIT grad student?",
-        "Can I attend HackMIT if I am a sophomore in high school?",
-        "I'm a high school student, but I'm really advanced. Can I attend HackMIT?",
-        "Do I need to bring money to the event?",
-        "Will we be able to sleep at the event?",
-        "Will we be able to stay overnight at the event?",
-        "What should I do if I am a beginner at the event?",
-    ]
-    results = query_all(3, questions)
-
-    for result in results:
-        print(result["query"])
-        for doc in result["result"]:
-            print(f"Score: {doc['score']}")
-            print(f"Source: {doc['source']}")
-            print(f"Q: {doc['question']}")
-            print(f"A: {doc['content']}")
-        print()
+# TODO(azliu): turn this into a test case
+# def test():
+#     try:
+#         embed_corpus()
+#     except Exception as err:
+#         print(f"Unexpected {err=}, {type(err)=}")
+#         raise
+
+#     questions = [
+#         "What is the deadline to apply for the hackathon?",
+#         "When is HackMIT?",
+#         "What are the challenges?",
+#         "How does judging work?",
+#         "What building should I go to during the event?",
+#         "What prizes are available?",
+#         "How many people are allowed on a team?",
+#         "What is HackMIT?",
+#         "Can I attend HackMIT if I am an MIT grad student?",
+#         "Can I attend HackMIT if I am a sophomore in high school?",
+#         "I'm a high school student, but I'm really advanced. Can I attend HackMIT?",
+#         "Do I need to bring money to the event?",
+#         "Will we be able to sleep at the event?",
+#         "Will we be able to stay overnight at the event?",
+#         "What should I do if I am a beginner at the event?",
+#     ]
+#     results = query_all(3, questions)
+
+#     for result in results:
+#         print(result["query"])
+#         for doc in result["result"]:
+#             print(f"Score: {doc['score']}")
+#             print(f"Source: {doc['source']}")
+#             print(f"Q: {doc['question']}")
+#             print(f"A: {doc['content']}")
+#         print()

server/nlp/responses.py

@@ -1,4 +1,4 @@
-"""Provides responses module.
+"""Responses.
 
 This module is used to generate responses to incoming emails using OpenAI.
 """