wip

aider/coders/base_coder.py

@@ -1,8 +1,10 @@
 #!/usr/bin/env python
 
+import json
 import os
 import sys
 import traceback
+from json.decoder import JSONDecodeError
 from pathlib import Path
 
 import backoff
@@ -356,7 +358,7 @@ class Coder:
             utils.show_messages(messages)
 
         try:
-            content, interrupted = self.send(messages)
+            interrupted = self.send(messages)
         except ExhaustedContextWindow:
             self.io.tool_error("Exhausted context window!")
             self.io.tool_error(" - Use /tokens to see token usage.")
@@ -364,6 +366,8 @@ class Coder:
             self.io.tool_error(" - Use /clear to clear chat history.")
             return
 
+        content = self.partial_response_content
+
         if interrupted:
             self.io.tool_error("\n\n^C KeyboardInterrupt")
             content += "\n^C KeyboardInterrupt"
@@ -468,19 +472,13 @@ class Coder:
         res = openai.ChatCompletion.create(**kwargs)
         return res
 
-        dump(res)
-        msg = res.choices[0].message
-        dump(msg)
-        print(msg.content)
-        print(msg.function_call.arguments)
-        sys.exit()
-        return res
-
     def send(self, messages, model=None, silent=False):
         if not model:
             model = self.main_model.name
 
-        self.resp = ""
+        self.partial_response_content = ""
+        self.partial_response_function_call = dict()
+
         interrupted = False
         try:
             completion = self.send_with_retries(model, messages)
@@ -489,9 +487,10 @@ class Coder:
             interrupted = True
 
         if not silent:
-            self.io.ai_output(self.resp)
+            # TODO: self.partial_response_function_call
+            self.io.ai_output(self.partial_response_content)
 
-        return self.resp, interrupted
+        return interrupted
 
     def show_send_output(self, completion, silent):
         live = None
@@ -503,20 +502,29 @@ class Coder:
                 live.start()
 
             for chunk in completion:
-                dump(chunk)
                 if chunk.choices[0].finish_reason == "length":
                     raise ExhaustedContextWindow()
 
                 try:
                     func = chunk.choices[0].delta.function_call
-                    print(func)
+                    # dump(func)
+                    for k, v in func.items():
+                        if k in self.partial_response_function_call:
+                            self.partial_response_function_call[k] += v
+                        else:
+                            self.partial_response_function_call[k] = v
+                    # dump(self.partial_response_function_call)
+                    args = self.partial_response_function_call.get("arguments")
+                    args = parse_partial_args(args)
+                    dump(args)
+
                 except AttributeError:
                     pass
 
                 try:
                     text = chunk.choices[0].delta.content
                     if text:
-                        self.resp += text
+                        self.partial_response_content += text
                 except AttributeError:
                     pass
 
@@ -524,7 +532,7 @@ class Coder:
                     continue
 
                 if self.pretty:
-                    show_resp = self.modify_incremental_response(self.resp)
+                    show_resp = self.modify_incremental_response()
                     md = Markdown(
                         show_resp, style=self.assistant_output_color, code_theme="default"
                     )
@@ -536,8 +544,8 @@ class Coder:
             if live:
                 live.stop()
 
-    def modify_incremental_response(self, resp):
-        return resp
+    def modify_incremental_response(self):
+        return self.partial_response_content
 
     def get_context_from_history(self, history):
         context = ""
@@ -562,7 +570,7 @@ class Coder:
         ]
 
         try:
-            commit_message, interrupted = self.send(
+            interrupted = self.send(
                 messages,
                 model=models.GPT35.name,
                 silent=True,
@@ -574,6 +582,7 @@ class Coder:
             )
             return
 
+        commit_message = self.partial_response_content
         commit_message = commit_message.strip()
         if commit_message and commit_message[0] == '"' and commit_message[-1] == '"':
             commit_message = commit_message[1:-1].strip()
@@ -728,3 +737,20 @@ def check_model_availability(main_model):
     available_models = openai.Model.list()
     model_ids = [model.id for model in available_models["data"]]
     return main_model.name in model_ids
+
+
+def parse_partial_args(data):
+    try:
+        return json.loads(data)
+    except JSONDecodeError:
+        pass
+
+    try:
+        return json.loads(data + "}")
+    except JSONDecodeError:
+        pass
+
+    try:
+        return json.loads(data + '"}')
+    except JSONDecodeError:
+        pass

aider/coders/func_coder.py

@@ -47,7 +47,8 @@ class FunctionCoder(Coder):
         else:
             self.cur_messages += [dict(role="assistant", content=content)]
 
-    def modify_incremental_response(self, resp):
+    def modify_incremental_response(self):
+        resp = self.partial_response_content
         return self.update_files(resp, mode="diff")
 
     def update_files(self, content, mode="update"):

aider/coders/wholefile_coder.py

@@ -20,7 +20,8 @@ class WholeFileCoder(Coder):
         else:
             self.cur_messages += [dict(role="assistant", content=content)]
 
-    def modify_incremental_response(self, resp):
+    def modify_incremental_response(self):
+        resp = self.partial_response_content
         return self.update_files(resp, mode="diff")
 
     def update_files(self, content, mode="update"):