Refactored code, added console output for syntax highlighting, and sk…

…ipped unknown programming languages in code parsing.
fynnfluegge · Sep 26, 2023 · 6f00743 · 6f00743
1 parent 6879e39
commit 6f00743
Show file tree

Hide file tree

Showing 4 changed files with 14 additions and 13 deletions.
diff --git a/codeqai/app.py b/codeqai/app.py
@@ -8,8 +8,7 @@
 from yaspin import yaspin
 
 from codeqai import codeparser, repo, utils
-from codeqai.config import (create_cache_dir, create_config, get_cache_path,
-                            load_config)
+from codeqai.config import create_cache_dir, create_config, get_cache_path, load_config
 from codeqai.constants import EmbeddingsModel, LllmHost
 from codeqai.embeddings import Embeddings
 from codeqai.llm import LLM
@@ -42,7 +41,7 @@ def run():
     create_cache_dir()
 
     embeddings_model = Embeddings(
-        local=True,
+        local=config["local"],
         model=EmbeddingsModel[config["embeddings"].upper().replace("-", "_")],
         deployment=config["embeddings-deployment"]
         if "embeddings-deployment" in config
@@ -94,11 +93,17 @@ def run():
                 language = utils.get_programming_language(
                     utils.get_file_extension(doc.metadata["filename"])
                 )
+
                 syntax = Syntax(
-                    doc.page_content, language.value, theme="monokai", line_numbers=True
+                    doc.page_content,
+                    language.value,
+                    theme="monokai",
+                    line_numbers=True,
                 )
                 console = Console()
+                print(doc.metadata["filename"] + " -> " + doc.metadata["method_name"])
                 console.print(syntax)
+                print()
 
             choice = input("[?] (C)ontinue search or (E)xit [C]:").strip().lower()
 

diff --git a/codeqai/codeparser.py b/codeqai/codeparser.py
@@ -1,5 +1,4 @@
 import os
-import re
 
 from langchain.schema import Document
 from langchain.text_splitter import RecursiveCharacterTextSplitter
@@ -19,9 +18,7 @@ def parse_code_files(code_files: list[str]) -> list[Document]:
             file_extension = utils.get_file_extension(code_file)
             programming_language = utils.get_programming_language(file_extension)
             if programming_language == Language.UNKNOWN:
-                print(
-                    f"Skipping file {code_file} with unsupported programming language"
-                )
+                continue
 
             langchain_language = utils.get_langchain_language(programming_language)
 

diff --git a/codeqai/config.py b/codeqai/config.py
@@ -100,7 +100,8 @@ def create_config():
                 "llm-host",
                 message="Which remote LLM do you want to use?",
                 choices=[
-                    "OpenAI" "Azure-OpenAI",
+                    "OpenAI",
+                    "Azure-OpenAI",
                 ],
                 default="OpenAI",
             ),

diff --git a/codeqai/embeddings.py b/codeqai/embeddings.py
@@ -17,12 +17,10 @@ def __init__(
         if not local:
             if model == EmbeddingsModel.OPENAI_TEXT_EMBEDDING_ADA_002:
                 self.embeddings = OpenAIEmbeddings(
-                    client=None, model="text_embedding_ada_002"
+                    client=None, model="text-embedding-ada-002"
                 )
             elif model == EmbeddingsModel.AZURE_OPENAI and deployment:
-                self.embeddings = OpenAIEmbeddings(
-                    client=None, model="text_embedding_ada_002", deployment=deployment
-                )
+                self.embeddings = OpenAIEmbeddings(client=None, deployment=deployment)
         else:
             try:
                 import sentence_transformers  # noqa: F401