chore: pr-review-tooling

codegen-examples/examples/pr_review_bot/run.py

@@ -1,7 +1,7 @@
 import codegen
 from codegen import Codebase
-from codegen.sdk.enums import ProgrammingLanguage
-from codegen.sdk.codebase.config import CodebaseConfig
+from codegen.shared.enums.programming_language import ProgrammingLanguage
+from codegen.sdk.codebase.config import CodebaseConfig, Secrets
 import json
 
 from codegen.sdk.secrets import Secrets
@@ -20,13 +20,9 @@ def run(codebase: Codebase):
     modified_symbols = codebase.get_modified_symbols_in_pr(pr_number)
     for symbol in modified_symbols:
         # Get direct dependencies
-        deps = codebase.get_symbol_dependencies(symbol, max_depth=2)
+        deps = symbol.dependencies(max_depth=2)
         context_symbols.update(deps)
 
-        # Get reverse dependencies (symbols that depend on this one)
-        rev_deps = codebase.get_symbol_dependents(symbol, max_depth=2)
-        context_symbols.update(rev_deps)
-
     # Prepare context for LLM
     context = {
         "modified_symbols": [

src/codegen/git/repo_operator/local_repo_operator.py

@@ -48,7 +48,7 @@
     ####################################################################################################################
 
     @property
     def remote_git_repo(self) -> GitRepoClient | None:
         """Get the remote GitRepoClient object for the current local repo."""
         if not self.access_token:
             msg = "Must initialize with access_token to get remote"
@@ -183,7 +183,7 @@
             if repo is None:
                 logger.warning("GitHub API key is required to fetch pull requests")
                 return None
-            return repo.get_pull(pr_number)
+            return repo.get_pull_safe(pr_number)
         except Exception as e:
             logger.warning(f"Failed to get PR {pr_number}: {e!s}")
             return None

src/codegen/git/repo_operator/repo_operator.py

@@ -65,7 +65,7 @@
     @property
     def remote_git_repo(self) -> GitRepoClient:
         if not self._remote_git_repo:
             self._remote_git_repo = GitRepoClient(self.repo_config, access_token=self.access_token)
         return self._remote_git_repo
 
     @property
@@ -111,7 +111,7 @@
         email_level = None
         levels = ["system", "global", "user", "repository"]
         for level in levels:
             with git_cli.config_reader(level) as reader:
                 if reader.has_option("user", "name") and not username:
                     username = reader.get("user", "name")
                     user_level = level
@@ -472,7 +472,7 @@
                     return content
             except UnicodeDecodeError:
                 print(f"Warning: Unable to decode file {file_path}. Skipping.")
                 return None
 
     def write_file(self, relpath: str, content: str) -> None:
         """Writes file content to disk"""
@@ -517,7 +517,7 @@
 
         # Iterate through files and yield contents
         for rel_filepath in filepaths:
             rel_filepath: str
             filepath = os.path.join(self.repo_path, rel_filepath)
 
             # Filter by subdirectory (includes full filenames)
@@ -548,7 +548,7 @@
         list_files = []
 
         for rel_filepath in self.git_cli.git.ls_files().split("\n"):
             rel_filepath: str
             if subdirs and not any(d in rel_filepath for d in subdirs):
                 continue
             if extensions is None or any(rel_filepath.endswith(e) for e in extensions):
@@ -572,7 +572,7 @@
 
     def get_modified_files_in_last_n_days(self, days: int = 1) -> tuple[list[str], list[str]]:
         """Returns a list of files modified and deleted in the last n days"""
         modified_files = []
         deleted_files = []
         allowed_extensions = [".py"]
 
@@ -588,9 +588,9 @@
                     if file in modified_files:
                         modified_files.remove(file)
                 else:
                     if file not in modified_files and file[-3:] in allowed_extensions:
                         modified_files.append(file)
         return modified_files, deleted_files
 
     @abstractmethod
     def base_url(self) -> str | None: ...
@@ -600,3 +600,11 @@
 
     def stash_pop(self) -> None:
         self.git_cli.git.stash("pop")
+
+    ####################################################################################################################
+    # PR UTILITIES
+    ####################################################################################################################
+
+    def get_pr_data(self, pr_number: int) -> dict:
+        """Returns the data associated with a PR"""
+        return self.remote_git_repo.get_pr_data(pr_number)

src/codegen/git/utils/pr_review.py

@@ -127,3 +127,20 @@ def modified_symbols(self) -> list["Symbol"]:
                 if self.is_modified(symbol):
                     all_modified.append(symbol)
         return all_modified
+
+    def get_pr_diff(self) -> str:
+        """Get the full diff of the PR"""
+        if not self._op.remote_git_repo:
+            msg = "GitHub API client is required to get PR diffs"
+            raise ValueError(msg)
+
+        # Get the diff directly from the PR
+        diff_url = self._gh_pr.raw_data.get("diff_url")
+        if diff_url:
+            # Fetch the diff content from the URL
+            response = requests.get(diff_url)
+            response.raise_for_status()
+            return response.text
+        else:
+            # If diff_url not available, get the patch directly
+            return self._gh_pr.get_patch()

src/codegen/sdk/core/codebase.py

@@ -1259,11 +1259,12 @@ def from_repo(
             logger.exception(f"Failed to initialize codebase: {e}")
             raise
 
-    def get_modified_symbols_in_pr(self, pr_id: int) -> list[Symbol]:
+    def get_modified_symbols_in_pr(self, pr_id: int) -> tuple[list[Symbol], str]:
         """Get all modified symbols in a pull request"""
         pr = self._op.get_pull_request(pr_id)
         cg_pr = CodegenPR(self._op, self, pr)
-        return cg_pr.modified_symbols
+        patch = cg_pr.get_pr_diff()
+        return cg_pr.modified_symbols, patch
 
 
 # The last 2 lines of code are added to the runner. See codegen-backend/cli/generate/utils.py