add --userguide argument

Author: Devin-Cook

aider/args.py

@@ -610,6 +610,12 @@ def get_parser(default_config_files, git_root):
         help="Run aider in your browser (default: False)",
         default=False,
     )
+    group.add_argument(
+        "--userguide",
+        action="store_true",
+        help="Use contect from ORE User Guide",
+        default=False,
+    )
     group.add_argument(
         "--copy-paste",
         action=argparse.BooleanOptionalAction,

aider/coders/base_coder.py

@@ -323,6 +323,7 @@ def __init__(
         file_watcher=None,
         auto_copy_context=False,
         auto_accept_architect=True,
+        userguide=False,
     ):
         # Fill in a dummy Analytics if needed, but it is never .enable()'d
         self.analytics = analytics if analytics is not None else Analytics()
@@ -409,6 +410,7 @@ def __init__(
 
         self.commands = commands or Commands(self.io, self)
         self.commands.coder = self
+        self.userguide = userguide
 
         self.repo = repo
         if use_git and self.repo is None:
@@ -519,6 +521,23 @@ def __init__(
                 self.io.tool_output("JSON Schema:")
                 self.io.tool_output(json.dumps(self.functions, indent=4))
 
+    def get_userguide_messages(self):
+        userguide_messages = []
+        query = self.cur_messages[-1]["content"]
+        userguide_messages.append(dict(
+            role="user",
+            content=("You may also use these *relevant sections of the ORE User Guide* so you can use them as additional context. If you reference these sections in your response, make sure to include section number.")
+        ))
+
+        from .userguide.query import query_message
+        userguide_messages.append((dict(role="user", content=query_message(query))))
+
+        userguide_messages.append(dict(
+            role="assistant",
+            content="Ok, I will use these sections as relevant context. If I mention these sections in my response, I will include the sectoin number."
+        ))
+        return userguide_messages
+    
     def setup_lint_cmds(self, lint_cmds):
         if not lint_cmds:
             return
@@ -1231,7 +1250,24 @@ def format_messages(self):
         chunks = self.format_chat_chunks()
         if self.add_cache_headers:
             chunks.add_cache_control_headers()
-
+        
+        # Handle userguide feature correctly - don't try to add lists to ChatChunks objects
+        if self.userguide:
+            try:
+                # Temporarily disable userguide feature to prevent errors
+                self.userguide = False
+                # Get the messages manually without using += operator
+                for msg in self.get_userguide_messages():
+                    if isinstance(chunks.cur, list):
+                        chunks.cur.append(msg)
+            except Exception as e:
+                # Log the error but continue
+                if hasattr(self, 'io'):
+                    self.io.tool_error(f"Error adding userguide messages: {e}")
+            finally:
+                # Restore the userguide setting
+                self.userguide = True
+        
         return chunks
 
     def warm_cache(self, chunks):

aider/coders/userguide/__init__.py

@@ -0,0 +1,107 @@
+import openai
+import os
+import pandas as pd
+import ast
+from scipy import spatial
+import tiktoken
+import logging
+
+# Set up logging
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger(__name__)
+
+# Try to get API key from environment
+api_key = os.environ.get("OPENAI_API_KEY")
+
+# Initialize OpenAI client only if API key is available
+client = None
+if api_key:
+    try:
+        client = openai.OpenAI(api_key=api_key)
+    except Exception as e:
+        logger.error(f"Failed to initialize OpenAI client: {e}")
+
+# Load embeddings from CSV
+try:
+    df = pd.read_csv("aider/aider/coders/userguide/userguide_embeddings.csv")
+    df['embedding'] = df['embedding'].apply(ast.literal_eval)
+except Exception as e:
+    logger.error(f"Failed to load embeddings: {e}")
+    # Create an empty DataFrame with required columns as fallback
+    df = pd.DataFrame(columns=["text", "embedding"])
+
+EMBEDDING_MODEL = "text-embedding-3-small"
+GPT_MODEL = "gpt-4o"
+
+def strings_ranked_by_relatedness(
+        query: str,
+        df: pd.DataFrame,
+        relatedness_fn=lambda x, y: 1 - spatial.distance.cosine(x, y),
+        top_n: int = 100
+) -> tuple[list[str], list[float]]:
+    """Returns a list of strings and relatedness, sorted from most related to least related."""
+    
+    # Check if client is available
+    if not client:
+        logger.warning("OpenAI client not initialized, returning empty results")
+        return [], []
+    
+    # Check if DataFrame is empty
+    if df.empty:
+        logger.warning("DataFrame is empty, returning empty results")
+        return [], []
+    
+    try:
+        # Fix the API call to match the new OpenAI client syntax
+        query_embedding_response = client.embeddings.create(
+            model=EMBEDDING_MODEL,
+            input=query  # Changed from 'inputs' to 'input'
+        )
+        query_embedding = query_embedding_response.data[0].embedding
+        
+        strings_and_relatedness = [
+            (row["text"], relatedness_fn(query_embedding, row["embedding"]))
+            for i, row in df.iterrows()
+        ]
+
+        strings_and_relatedness.sort(key=lambda x: x[1], reverse=True)
+        strings, relatedness = zip(*strings_and_relatedness)
+        return strings[:top_n], relatedness[:top_n]
+    except Exception as e:
+        logger.error(f"Error in relatedness calculation: {e}")
+        return [], []
+    
+def num_tokens(text: str, model: str = GPT_MODEL) -> int:
+    try:
+        encoding = tiktoken.encoding_for_model(model)
+        return len(encoding.encode(text))
+    except Exception:
+        # Fallback token estimation
+        return len(text) // 4
+
+def query_message(
+        query: str,
+        df: pd.DataFrame = df, 
+        model: str = GPT_MODEL, 
+        token_budget: int = 4096-500
+) -> str:
+    """Return a message for GPT that, with relevant source texts pulled from a dataframe."""
+    try:
+        strings, relatedness = strings_ranked_by_relatedness(query, df)
+        
+        if not strings:
+            return "Unable to retrieve relevant sections from the user guide."
+            
+        message = ""
+        for string in strings:
+            next_article = f'\n\nUser Guide section:\n"""\n{string}\n"""'
+            if (
+                num_tokens(message + next_article, model) > token_budget
+            ):
+                break
+            else:
+                message += next_article
+        return message
+    except Exception as e:
+        logger.error(f"Error in query_message: {e}")
+        return "An error occurred while searching the user guide. Proceeding without user guide references."

aider/coders/userguide/query.py

@@ -38,7 +38,7 @@ def get_captured_lines(self):
 
 def search(text=None):
     results = []
-    for root, _, files in os.walk("aider"):
+    for root, _, files in os.walk("explORE-AI"):
         for file in files:
             path = os.path.join(root, file)
             if not text or text in path:
@@ -148,20 +148,13 @@ def add_undo(self, commit_hash):
 
     def do_sidebar(self):
         with st.sidebar:
-            st.title("Aider")
+            st.image(urls.ORE_Image, caption=None, width=None, use_column_width=None, clamp=False, channels="RGB", output_format="auto", use_container_width=False)
+            st.title("explORE-AI")
             # self.cmds_tab, self.settings_tab = st.tabs(["Commands", "Settings"])
-
             # self.do_recommended_actions()
             self.do_add_to_chat()
             self.do_recent_msgs()
             self.do_clear_chat_history()
-            # st.container(height=150, border=False)
-            # st.write("### Experimental")
-
-            st.warning(
-                "This browser version of aider is experimental. Please share feedback in [GitHub"
-                " issues](https://github.com/Aider-AI/aider/issues)."
-            )
 
     def do_settings_tab(self):
         pass
@@ -182,7 +175,7 @@ def do_recommended_actions(self):
     def do_add_to_chat(self):
         # with st.expander("Add to the chat", expanded=True):
         self.do_add_files()
-        self.do_add_web_page()
+        # self.do_add_web_page()
 
     def do_add_files(self):
         fnames = st.multiselect(
@@ -191,10 +184,10 @@ def do_add_files(self):
             default=self.state.initial_inchat_files,
             placeholder="Files to edit",
             disabled=self.prompt_pending(),
-            help=(
-                "Only add the files that need to be *edited* for the task you are working"
-                " on. Aider will pull in other relevant code to provide context to the LLM."
-            ),
+            # help=(
+            #     "Only add the files that need to be *edited* for the task you are working"
+            #     " on. Aider will pull in other relevant code to provide context to the LLM."
+            # ),
         )
 
         for fname in fnames:
@@ -318,10 +311,12 @@ def do_messages_container(self):
                     line = text.splitlines()[0]
                     with self.messages.expander(line):
                         st.text(text)
-                elif role in ("user", "assistant"):
-                    with st.chat_message(role):
+                elif role == "user":
+                    with st.chat_message("user"):
+                        st.write(msg["content"])
+                elif role == "assistant":
+                    with st.chat_message("assistant", avatar=urls.ORE_Favicon):
                         st.write(msg["content"])
-                        # self.cost()
                 else:
                     st.dict(msg)
 
@@ -418,7 +413,7 @@ def process_chat(self):
         self.max_reflections = 3
 
         while prompt:
-            with self.messages.chat_message("assistant"):
+            with self.messages.chat_message("assistant", avatar=urls.ORE_Favicon):
                 res = st.write_stream(self.coder.run_stream(prompt))
                 self.state.messages.append({"role": "assistant", "content": res})
                 # self.cost()
@@ -524,12 +519,12 @@ def do_undo(self, commit_hash):
 def gui_main():
     st.set_page_config(
         layout="wide",
-        page_title="Aider",
-        page_icon=urls.favicon,
+        page_title="explORE-AI",
+        page_icon=urls.ORE_Favicon,
         menu_items={
-            "Get Help": urls.website,
-            "Report a bug": "https://github.com/Aider-AI/aider/issues",
-            "About": "# Aider\nAI pair programming in your browser.",
+            "Get Help": urls.ORE_GitHub,
+            # "Report a bug": "https://github.com/Aider-AI/aider/issues",
+            "About": "# explORE-AI assistant for understanding the ORE codebase", 
         },
     )
 

aider/coders/userguide/userguide_embeddings.csv

@@ -948,6 +948,7 @@ def get_io(pretty):
             detect_urls=args.detect_urls,
             auto_copy_context=args.copy_paste,
             auto_accept_architect=args.auto_accept_architect,
+            userguide=args.userguide,
         )
     except UnknownEditFormat as err:
         io.tool_error(str(err))

aider/gui.py

@@ -30,7 +30,7 @@
 )
 
 added_files = (
-    "I added these files to the chat: {fnames}\nLet me know if there are others we should add."
+    "I added these files to the chat: {fnames}\nLet me know if there are others we should add. If you need more context from other existing files not already added to the chat, you *MUST* tell me their full path names and ask me to *add the files to the chat* end your reply and wait for my approval. you can keep asking if you then dicide you need to view more files."
 )
 
 

aider/main.py

@@ -15,3 +15,6 @@
 release_notes = "https://aider.chat/HISTORY.html#release-notes"
 edit_formats = "https://aider.chat/docs/more/edit-formats.html"
 models_and_keys = "https://aider.chat/docs/troubleshooting/models-and-keys.html"
+ORE_Favicon = "https://www.opensourcerisk.org/wp-content/uploads/2016/10/favicon.png"
+ORE_Image = "https://www.opensourcerisk.org/wp-content/uploads/2015/03/logo_trans-1.png"
+ORE_GitHub = "https://github.com/OpenSourceRisk/Engine"