Merge pull request Codium-ai#159 from Codium-ai/tr/edit_any_config_se…

…tting

The Configurator Strikes Back

CHANGELOG.md

@@ -1,3 +1,10 @@
+## 2023-07-30
+
+### Enhanced
+- Added the ability to modify any configuration parameter from 'configuration.toml' on-the-fly.
+- Updated the command line interface and bot commands to accept configuration changes as arguments.
+- Improved the PR agent to handle additional arguments for each action.
+
 ## 2023-07-28
 
 ### Improved

pr_agent/agent/pr_agent.py

@@ -16,16 +16,16 @@ def __init__(self):
     async def handle_request(self, pr_url, request) -> bool:
         action, *args = request.strip().split()
         if any(cmd == action for cmd in ["/answer"]):
-            await PRReviewer(pr_url, is_answer=True).review()
+            await PRReviewer(pr_url, is_answer=True, args=args).review()
         elif any(cmd == action for cmd in ["/review", "/review_pr", "/reflect_and_review"]):
             if settings.pr_reviewer.ask_and_reflect or "/reflect_and_review" in request:
-                await PRInformationFromUser(pr_url).generate_questions()
+                await PRInformationFromUser(pr_url, args=args).generate_questions()
             else:
                 await PRReviewer(pr_url, args=args).review()
         elif any(cmd == action for cmd in ["/describe", "/describe_pr"]):
             await PRDescription(pr_url, args=args).describe()
         elif any(cmd == action for cmd in ["/improve", "/improve_code"]):
-            await PRCodeSuggestions(pr_url).suggest()
+            await PRCodeSuggestions(pr_url, args=args).suggest()
         elif any(cmd == action for cmd in ["/ask", "/ask_question"]):
             await PRQuestions(pr_url, args=args).answer()
         elif any(cmd == action for cmd in ["/update_changelog"]):

pr_agent/algo/utils.py

@@ -1,4 +1,5 @@
 from __future__ import annotations
+from typing import List
 
 import difflib
 from datetime import datetime
@@ -211,3 +212,46 @@ def load_large_diff(file, new_file_content_str: str, original_file_content_str:
         except Exception:
             pass
     return patch
+
+
+def update_settings_from_args(args: List[str]) -> None:
+    """
+    Update the settings of the Dynaconf object based on the arguments passed to the function.
+
+    Args:
+        args: A list of arguments passed to the function.
+        Example args: ['--pr_code_suggestions.extra_instructions="be funny',
+                  '--pr_code_suggestions.num_code_suggestions=3']
+
+    Returns:
+        None
+
+    Raises:
+        ValueError: If the argument is not in the correct format.
+
+    """
+    if args:
+        for arg in args:
+            try:
+                arg = arg.strip('-').strip()
+                vals = arg.split('=')
+                if len(vals) != 2:
+                    raise ValueError(f'Invalid argument format: {arg}')
+                key, value = vals
+                keys = key.split('.')
+                d = settings
+                for i, k in enumerate(keys[:-1]):
+                    if k not in d:
+                        raise ValueError(f'Invalid setting: {key}')
+                    d = d[k]
+                if keys[-1] not in d:
+                    raise ValueError(f'Invalid setting: {key}')
+                if isinstance(d[keys[-1]], bool):
+                    d[keys[-1]] = value.lower() in ("yes", "true", "t", "1")
+                else:
+                    d[keys[-1]] = type(d[keys[-1]])(value)
+                logging.info(f'Updated setting {key} to: "{value}"')
+            except ValueError as e:
+                logging.error(str(e))
+            except Exception as e:
+                logging.error(f'Failed to parse argument {arg}: {e}')

pr_agent/cli.py

@@ -29,6 +29,9 @@ def run(args=None):
 improve / improve_code - Suggest improvements to the code in the PR as pull request comments ready to commit.
 reflect - Ask the PR author questions about the PR.
 update_changelog - Update the changelog based on the PR's contents.
+
+To edit any configuration parameter from 'configuration.toml', just add -config_path=<value>.
+For example: '- cli.py --pr-url=... review --pr_reviewer.extra_instructions="focus on the file: ..."'
 """)
     parser.add_argument('--pr_url', type=str, help='The URL of the PR to review', required=True)
     parser.add_argument('command', type=str, help='The', choices=['review', 'review_pr',
@@ -79,7 +82,7 @@ def _handle_describe_command(pr_url: str, rest: list):
 
 def _handle_improve_command(pr_url: str, rest: list):
     print(f"PR code suggestions: {pr_url}")
-    reviewer = PRCodeSuggestions(pr_url)
+    reviewer = PRCodeSuggestions(pr_url, args=rest)
     asyncio.run(reviewer.suggest())
 
 
@@ -97,7 +100,7 @@ def _handle_reflect_command(pr_url: str, rest: list):
 
 def _handle_review_after_reflect_command(pr_url: str, rest: list):
     print(f"Processing author's answers and sending review: {pr_url}")
-    reviewer = PRReviewer(pr_url, cli_mode=True, is_answer=True)
+    reviewer = PRReviewer(pr_url, cli_mode=True, is_answer=True, args=rest)
     asyncio.run(reviewer.review())
 
 def _handle_update_changelog(pr_url: str, rest: list):

pr_agent/config_loader.py

@@ -19,7 +19,7 @@
          "settings/pr_description_prompts.toml",
          "settings/pr_code_suggestions_prompts.toml",
          "settings/pr_information_from_user_prompts.toml",
-         "settings/pr_update_changelog.toml",
+         "settings/pr_update_changelog_prompts.toml",
          "settings_prod/.secrets.toml"
         ]]
 )

pr_agent/servers/help.py

@@ -1,11 +1,10 @@
 commands_text = "> **/review [-i]**: Request a review of your Pull Request. For an incremental review, which only " \
                 "considers changes since the last review, include the '-i' option.\n" \
-                "> **/describe [-c]**: Modify the PR title and description based on the contents of the PR. " \
-                "To get the description as comment instead of modifying the PR description, " \
-                "include the '-c' option.\n" \
-                "> **/improve**: Suggest improvements to the code in the PR. " \
-                "These will be provided as pull request comments, ready to commit.\n" \
-                "> **/ask \\<QUESTION\\>**: Pose a question about the PR.\n"
+                "> **/describe**: Modify the PR title and description based on the contents of the PR.\n" \
+                "> **/improve**: Suggest improvements to the code in the PR. \n" \
+                "> **/ask \\<QUESTION\\>**: Pose a question about the PR.\n\n" \
+                "To edit any configuration parameter from 'configuration.toml', just add -config_path=<value>. " \
+                "For example: '/review --pr_reviewer.extra_instructions=\"focus on the file: ...\"'" \
 
 
 def bot_help_text(user: str):

pr_agent/settings/configuration.toml

@@ -7,25 +7,29 @@ publish_output_progress=true
 verbosity_level=0 # 0,1,2
 use_extra_bad_extensions=false
 
-[pr_reviewer]
+[pr_reviewer] # /review #
 require_focused_review=true
 require_score_review=false
 require_tests_review=true
 require_security_review=true
 num_code_suggestions=0
 inline_code_comments = true
 ask_and_reflect=false
+extra_instructions = ""
 
-[pr_description]
+[pr_description] # /describe #
 publish_description_as_comment=false
+extra_instructions = ""
 
-[pr_questions]
+[pr_questions] # /ask #
 
-[pr_code_suggestions]
+[pr_code_suggestions] # /improve #
 num_code_suggestions=4
+extra_instructions = ""
 
-[pr_update_changelog]
+[pr_update_changelog] # /update_changelog #
 push_changelog_changes=false
+extra_instructions = ""
 
 [github]
 # The type of deployment to create. Valid values are 'app' or 'user'.

pr_agent/settings/pr_code_suggestions_prompts.toml

@@ -9,6 +9,12 @@ Your task is to provide meaningfull non-trivial code suggestions to improve the
 - Make sure not to provide suggestions repeating modifications already implemented in the new PR code (the '+' lines).
 - Don't output line numbers in the 'improved code' snippets.
 
+{%- if extra_instructions %}
+
+Extra instructions from the user:
+{{ extra_instructions }}
+{% endif %}
+
 You must use the following JSON schema to format your answer:
 ```json
 {

pr_agent/settings/pr_description_prompts.toml

@@ -3,6 +3,12 @@ system="""You are CodiumAI-PR-Reviewer, a language model designed to review git
 Your task is to provide full description of the PR content.
 - Make sure not to focus the new PR code (the '+' lines).
 
+{%- if extra_instructions %}
+
+Extra instructions from the user:
+{{ extra_instructions }}
+{% endif %}
+
 You must use the following JSON schema to format your answer:
 ```json
 {

pr_agent/settings/pr_reviewer_prompts.toml

@@ -8,6 +8,12 @@ Your task is to provide constructive and concise feedback for the PR, and also p
 - Make sure not to provide suggestions repeating modifications already implemented in the new PR code (the '+' lines).
 {%- endif %}
 
+{%- if extra_instructions %}
+
+Extra instructions from the user:
+{{ extra_instructions }}
+{% endif %}
+
 You must use the following JSON schema to format your answer:
 ```json
 {

pr_agent/settings/pr_update_changelog.toml → pr_agent/settings/pr_update_changelog_prompts.toml

@@ -4,6 +4,12 @@ Your task is to update the CHANGELOG.md file of the project, to shortly summariz
 - The output should match the existing CHANGELOG.md format, style and conventions, so it will look like a natural part of the file. For example, if previous changes were summarized in a single line, you should do the same.
 - Don't repeat previous changes. Generate only new content, that is not already in the CHANGELOG.md file.
 - Be general, and avoid specific details, files, etc. The output should be minimal, no more than 3-4 short lines. Ignore non-relevant subsections.
+
+{%- if extra_instructions %}
+
+Extra instructions from the user:
+{{ extra_instructions }}
+{%- endif %}
 """
 
 user="""PR Info:

pr_agent/tools/pr_code_suggestions.py

@@ -8,19 +8,21 @@
 from pr_agent.algo.ai_handler import AiHandler
 from pr_agent.algo.pr_processing import get_pr_diff, retry_with_fallback_models
 from pr_agent.algo.token_handler import TokenHandler
-from pr_agent.algo.utils import try_fix_json
+from pr_agent.algo.utils import try_fix_json, update_settings_from_args
 from pr_agent.config_loader import settings
 from pr_agent.git_providers import BitbucketProvider, get_git_provider
 from pr_agent.git_providers.git_provider import get_main_pr_language
 
 
 class PRCodeSuggestions:
-    def __init__(self, pr_url: str, cli_mode=False):
+    def __init__(self, pr_url: str, cli_mode=False, args: list = None):
 
         self.git_provider = get_git_provider()(pr_url)
         self.main_language = get_main_pr_language(
             self.git_provider.get_languages(), self.git_provider.get_files()
         )
+        update_settings_from_args(args)
+
         self.ai_handler = AiHandler()
         self.patches_diff = None
         self.prediction = None
@@ -31,7 +33,8 @@ def __init__(self, pr_url: str, cli_mode=False):
             "description": self.git_provider.get_pr_description(),
             "language": self.main_language,
             "diff": "",  # empty diff for initial calculation
-            'num_code_suggestions': settings.pr_code_suggestions.num_code_suggestions,
+            "num_code_suggestions": settings.pr_code_suggestions.num_code_suggestions,
+            "extra_instructions": settings.pr_code_suggestions.extra_instructions,
         }
         self.token_handler = TokenHandler(self.git_provider.pr,
                                           self.vars,
@@ -137,3 +140,4 @@ def dedent_code(self, relevant_file, relevant_lines_start, new_code_snippet):
                 logging.info(f"Could not dedent code snippet for file {relevant_file}, error: {e}")
 
         return new_code_snippet
+

pr_agent/tools/pr_description.py

@@ -8,6 +8,7 @@
 from pr_agent.algo.ai_handler import AiHandler
 from pr_agent.algo.pr_processing import get_pr_diff, retry_with_fallback_models
 from pr_agent.algo.token_handler import TokenHandler
+from pr_agent.algo.utils import update_settings_from_args
 from pr_agent.config_loader import settings
 from pr_agent.git_providers import get_git_provider
 from pr_agent.git_providers.git_provider import get_main_pr_language
@@ -21,8 +22,8 @@ def __init__(self, pr_url: str, args: list = None):
             pr_url (str): The URL of the pull request.
             args (list, optional): List of arguments passed to the PRDescription class. Defaults to None.
         """
-        self.parse_args(args)
-        
+        update_settings_from_args(args)
+
         # Initialize the git provider and main PR language
         self.git_provider = get_git_provider()(pr_url)
         self.main_pr_language = get_main_pr_language(
@@ -39,6 +40,7 @@ def __init__(self, pr_url: str, args: list = None):
             "description": self.git_provider.get_pr_description(),
             "language": self.main_pr_language,
             "diff": "",  # empty diff for initial calculation
+            "extra_instructions": settings.pr_description.extra_instructions,
         }
 
         # Initialize the token handler
@@ -53,22 +55,6 @@ def __init__(self, pr_url: str, args: list = None):
         self.patches_diff = None
         self.prediction = None
 
-    def parse_args(self, args: List[str]) -> None:
-        """
-        Parse the arguments passed to the PRDescription class and set the 'publish_description_as_comment' attribute accordingly.
-
-        Args:
-            args: A list of arguments passed to the PRReviewer class.
-
-        Returns:
-            None
-        """
-        self.publish_description_as_comment = settings.pr_description.publish_description_as_comment
-        if args and len(args) >= 1:
-            arg = args[0]
-            if arg == "-c":
-                self.publish_description_as_comment = True
-
     async def describe(self):
         """
         Generates a PR description using an AI model and publishes it to the PR.
@@ -84,7 +70,7 @@ async def describe(self):
 
         if settings.config.publish_output:
             logging.info('Pushing answer...')
-            if self.publish_description_as_comment:
+            if settings.pr_description.publish_description_as_comment:
                 self.git_provider.publish_comment(markdown_text)
             else:
                 self.git_provider.publish_description(pr_title, pr_body)

pr_agent/tools/pr_information_from_user.py

@@ -14,7 +14,7 @@
 
 
 class PRInformationFromUser:
-    def __init__(self, pr_url: str):
+    def __init__(self, pr_url: str, args: list = None):
         self.git_provider = get_git_provider()(pr_url)
         self.main_pr_language = get_main_pr_language(
             self.git_provider.get_languages(), self.git_provider.get_files()

pr_agent/tools/pr_reviewer.py

@@ -9,7 +9,7 @@
 from pr_agent.algo.ai_handler import AiHandler
 from pr_agent.algo.pr_processing import get_pr_diff, retry_with_fallback_models
 from pr_agent.algo.token_handler import TokenHandler
-from pr_agent.algo.utils import convert_to_markdown, try_fix_json
+from pr_agent.algo.utils import convert_to_markdown, try_fix_json, update_settings_from_args
 from pr_agent.config_loader import settings
 from pr_agent.git_providers import get_git_provider
 from pr_agent.git_providers.git_provider import get_main_pr_language, IncrementalPR
@@ -30,7 +30,8 @@ def __init__(self, pr_url: str, cli_mode: bool = False, is_answer: bool = False,
             is_answer (bool, optional): Indicates whether the review is being done in answer mode. Defaults to False.
             args (list, optional): List of arguments passed to the PRReviewer class. Defaults to None.
         """
-        self.parse_args(args)
+        update_settings_from_args(args)
+        self.parse_args(args) # -i command
 
         self.git_provider = get_git_provider()(pr_url, incremental=self.incremental)
         self.main_language = get_main_pr_language(
@@ -60,6 +61,7 @@ def __init__(self, pr_url: str, cli_mode: bool = False, is_answer: bool = False,
             'num_code_suggestions': settings.pr_reviewer.num_code_suggestions,
             'question_str': question_str,
             'answer_str': answer_str,
+            "extra_instructions": settings.pr_reviewer.extra_instructions,
         }
 
         self.token_handler = TokenHandler(