ai-cr 3.2.2__py3-none-any.whl → 3.3.0__py3-none-any.whl

gito/core.py

@@ -1,508 +1,520 @@
-import os
-import fnmatch
-import logging
-from typing import Iterable
-from pathlib import Path
-from functools import partial
-
-import microcore as mc
-from gito.gh_api import gh_api
-from microcore import ui
-from git import Repo, Commit
-from git.exc import GitCommandError
-from unidiff import PatchSet, PatchedFile
-from unidiff.constants import DEV_NULL
-
-from .context import Context
-from .project_config import ProjectConfig
-from .report_struct import Report
-from .constants import JSON_REPORT_FILE_NAME
-from .utils import make_streaming_function
-from .pipeline import Pipeline
-from .env import Env
-
-
-def review_subject_is_index(what):
-    return not what or what == 'INDEX'
-
-
-def is_binary_file(repo: Repo, file_path: str) -> bool:
-    """
-    Check if a file is binary by attempting to read it as text.
-    Returns True if the file is binary, False otherwise.
-    """
-    try:
-        # Attempt to read the file content from the repository tree
-        content = repo.tree()[file_path].data_stream.read()
-        # Try decoding as UTF-8; if it fails, it's likely binary
-        content.decode("utf-8")
-        return False
-    except KeyError:
-        try:
-            fs_path = Path(repo.working_tree_dir) / file_path
-            fs_path.read_text(encoding='utf-8')
-            return False
-        except FileNotFoundError:
-            logging.error(f"File {file_path} not found in the repository.")
-            return True
-        except UnicodeDecodeError:
-            return True
-        except Exception as e:
-            logging.error(f"Error reading file {file_path}: {e}")
-            return True
-    except UnicodeDecodeError:
-        return True
-    except Exception as e:
-        logging.warning(f"Error checking if file {file_path} is binary: {e}")
-        return True  # Conservatively treat errors as binary to avoid issues
-
-
-def commit_in_branch(repo: Repo, commit: Commit, target_branch: str) -> bool:
-    try:
-        # exit code 0 if commit is ancestor of branch
-        repo.git.merge_base('--is-ancestor', commit.hexsha, target_branch)
-        return True
-    except GitCommandError:
-        pass
-    return False
-
-
-def get_base_branch(repo: Repo, pr: int | str = None):
-    if os.getenv('GITHUB_ACTIONS'):
-
-        # triggered from PR
-        if base_ref := os.getenv('GITHUB_BASE_REF'):
-            logging.info(f"Using GITHUB_BASE_REF:{base_ref} as base branch")
-            return f'origin/{base_ref}'
-        logging.info("GITHUB_BASE_REF is not available...")
-        if pr:
-            api = gh_api(repo=repo)
-            pr_obj = api.pulls.get(pr)
-            logging.info(
-                f"Using 'origin/{pr_obj.base.ref}' as base branch "
-                f"(received via GH API for PR#{pr})"
-            )
-            return f'origin/{pr_obj.base.ref}'
-
-    try:
-        logging.info(
-            "Trying to resolve base branch from repo.remotes.origin.refs.HEAD.reference.name..."
-        )
-        # 'origin/main', 'origin/master', etc
-        # Stopped working in github actions since 07/2025
-        return repo.remotes.origin.refs.HEAD.reference.name
-    except AttributeError:
-        try:
-            logging.info(
-                "Checking if repo has 'main' or 'master' branchs to use as --against branch..."
-            )
-            remote_refs = repo.remotes.origin.refs
-            for branch_name in ['main', 'master']:
-                if hasattr(remote_refs, branch_name):
-                    return f'origin/{branch_name}'
-        except Exception:
-            pass
-
-        logging.error("Could not determine default branch from remote refs.")
-        raise ValueError("No default branch found in the repository.")
-
-
-def get_diff(
-    repo: Repo = None,
-    what: str = None,
-    against: str = None,
-    use_merge_base: bool = True,
-    pr: str | int = None
-) -> PatchSet | list[PatchedFile]:
-    repo = repo or Repo(".")
-    if not against:
-        # 'origin/main', 'origin/master', etc
-        against = get_base_branch(repo, pr=pr)
-    if review_subject_is_index(what):
-        what = None  # working copy
-    if use_merge_base:
-        try:
-            if review_subject_is_index(what):
-                try:
-                    current_ref = repo.active_branch.name
-                except TypeError:
-                    # In detached HEAD state, use HEAD directly
-                    current_ref = "HEAD"
-                    logging.info(
-                        "Detected detached HEAD state, using HEAD as current reference"
-                    )
-            else:
-                current_ref = what
-            merge_base = repo.merge_base(current_ref or repo.active_branch.name, against)[0]
-            logging.info(
-                f"Merge base({ui.green(current_ref)},{ui.yellow(against)})"
-                f" --> {ui.cyan(merge_base.hexsha)}"
-            )
-            # if branch is already an ancestor of "against", merge_base == branch ⇒ it’s been merged
-            if merge_base.hexsha == repo.commit(current_ref or repo.active_branch.name).hexsha:
-                # @todo: check case: reviewing working copy index in main branch #103
-                logging.info(
-                    f"Branch is already merged. ({ui.green(current_ref)} vs {ui.yellow(against)})"
-                )
-                merge_sha = repo.git.log(
-                    '--merges',
-                    '--ancestry-path',
-                    f'{current_ref}..{against}',
-                    '-n',
-                    '1',
-                    '--pretty=format:%H'
-                ).strip()
-                if merge_sha:
-                    logging.info(f"Merge commit is {ui.cyan(merge_sha)}")
-                    merge_commit = repo.commit(merge_sha)
-
-                    other_merge_parent = None
-                    for parent in merge_commit.parents:
-                        logging.info(f"Checking merge parent: {parent.hexsha[:8]}")
-                        if parent.hexsha == merge_base.hexsha:
-                            logging.info(f"merge parent is {ui.cyan(parent.hexsha[:8])}, skipping")
-                            continue
-                        if not commit_in_branch(repo, parent, against):
-                            logging.warning(f"merge parent is not in {against}, skipping")
-                            continue
-                        logging.info(f"Found other merge parent: {ui.cyan(parent.hexsha[:8])}")
-                        other_merge_parent = parent
-                        break
-                    if other_merge_parent:
-                        first_common_ancestor = repo.merge_base(other_merge_parent, merge_base)[0]
-                        # for gito remote (feature_branch vs origin/main)
-                        # the same merge base appears in first_common_ancestor again
-                        if first_common_ancestor.hexsha == merge_base.hexsha:
-                            if merge_base.parents:
-                                first_common_ancestor = repo.merge_base(
-                                    other_merge_parent, merge_base.parents[0]
-                                )[0]
-                            else:
-                                logging.error(
-                                    "merge_base has no parents, "
-                                    "using merge_base as first_common_ancestor"
-                                )
-                        logging.info(
-                            f"{what} will be compared to "
-                            f"first common ancestor of {what} and {against}: "
-                            f"{ui.cyan(first_common_ancestor.hexsha[:8])}"
-                        )
-                        against = first_common_ancestor.hexsha
-                    else:
-                        logging.error(f"Can't find other merge parent for {merge_sha}")
-                else:
-                    logging.warning(
-                        f"No merge‐commit found for {current_ref!r}→{against!r}; "
-                        "falling back to merge‐base diff"
-                    )
-            else:
-                # normal case: branch not yet merged
-                against = merge_base.hexsha
-                logging.info(
-                    f"Using merge base: {ui.cyan(merge_base.hexsha[:8])} ({merge_base.summary})"
-                )
-        except Exception as e:
-            logging.error(f"Error finding merge base: {e}")
-    logging.info(
-        f"Making diff: {ui.green(what or 'INDEX')} vs {ui.yellow(against)}"
-    )
-    diff_content = repo.git.diff(against, what)
-    diff = PatchSet.from_string(diff_content)
-
-    # Filter out binary files
-    non_binary_diff = PatchSet([])
-    for patched_file in diff:
-        # Check if the file is binary using the source or target file path
-        file_path = (
-            patched_file.target_file
-            if patched_file.target_file != DEV_NULL
-            else patched_file.source_file
-        )
-        if file_path == DEV_NULL:
-            continue
-        if is_binary_file(repo, file_path.removeprefix("b/")):
-            logging.info(f"Skipping binary file: {patched_file.path}")
-            continue
-        non_binary_diff.append(patched_file)
-    return non_binary_diff
-
-
-def filter_diff(
-    patch_set: PatchSet | Iterable[PatchedFile], filters: str | list[str]
-) -> PatchSet | Iterable[PatchedFile]:
-    """
-    Filter the diff files by the given fnmatch filters.
-    """
-    assert isinstance(filters, (list, str))
-    if not isinstance(filters, list):
-        filters = [f.strip() for f in filters.split(",") if f.strip()]
-    if not filters:
-        return patch_set
-    files = [
-        file
-        for file in patch_set
-        if any(fnmatch.fnmatch(file.path, pattern) for pattern in filters)
-    ]
-    return files
-
-
-def read_file(repo: Repo, file: str, use_local_files: bool = False) -> str:
-    if use_local_files:
-        file_path = Path(repo.working_tree_dir) / file
-        try:
-            return file_path.read_text(encoding='utf-8')
-        except (FileNotFoundError, UnicodeDecodeError) as e:
-            logging.warning(f"Could not read file {file} from working directory: {e}")
-
-    # Read from HEAD (committed version)
-    return repo.tree()[file].data_stream.read().decode('utf-8')
-
-
-def file_lines(repo: Repo, file: str, max_tokens: int = None, use_local_files: bool = False) -> str:
-    text = read_file(repo=repo, file=file, use_local_files=use_local_files)
-    lines = [f"{i + 1}: {line}\n" for i, line in enumerate(text.splitlines())]
-    if max_tokens:
-        lines, removed_qty = mc.tokenizing.fit_to_token_size(lines, max_tokens)
-        if removed_qty:
-            lines.append(
-                f"(!) DISPLAYING ONLY FIRST {len(lines)} LINES DUE TO LARGE FILE SIZE\n"
-            )
-    return "".join(lines)
-
-
-def read_files(repo: Repo, files: list[str], max_tokens: int = None) -> dict:
-    out = dict()
-    total_tokens = 0
-    for file in files:
-        content = read_file(repo=repo, file=file, use_local_files=True)
-        total_tokens += mc.tokenizing.num_tokens_from_string(file)
-        total_tokens += mc.tokenizing.num_tokens_from_string(content)
-        if max_tokens and total_tokens > max_tokens:
-            logging.warning(
-                f"Skipping file {file} due to exceeding max_tokens limit ({max_tokens})"
-            )
-            continue
-        out[file] = content
-    return out
-
-
-def make_cr_summary(ctx: Context, **kwargs) -> str:
-    return (
-        mc.prompt(
-            ctx.config.summary_prompt,
-            diff=mc.tokenizing.fit_to_token_size(ctx.diff, ctx.config.max_code_tokens)[0],
-            issues=ctx.report.issues,
-            pipeline_out=ctx.pipeline_out,
-            env=Env,
-            **ctx.config.prompt_vars,
-            **kwargs,
-        ).to_llm()
-        if ctx.config.summary_prompt
-        else ""
-    )
-
-
-class NoChangesInContextError(Exception):
-    """
-    Exception raised when there are no changes in the context to review /answer questions.
-    """
-
-
-def _prepare(
-    repo: Repo = None,
-    what: str = None,
-    against: str = None,
-    filters: str | list[str] = "",
-    use_merge_base: bool = True,
-    pr: str | int = None,
-):
-    repo = repo or Repo(".")
-    cfg = ProjectConfig.load_for_repo(repo)
-    diff = get_diff(
-        repo=repo, what=what, against=against, use_merge_base=use_merge_base, pr=pr,
-    )
-    diff = filter_diff(diff, filters)
-    if not diff:
-        raise NoChangesInContextError()
-    lines = {
-        file_diff.path: (
-            file_lines(
-                repo,
-                file_diff.path,
-                cfg.max_code_tokens
-                - mc.tokenizing.num_tokens_from_string(str(file_diff)),
-                use_local_files=review_subject_is_index(what)
-            )
-            if file_diff.target_file != DEV_NULL and not file_diff.is_added_file
-            else ""
-        )
-        for file_diff in diff
-    }
-    return repo, cfg, diff, lines
-
-
-def get_affected_code_block(repo: Repo, file: str, start_line: int, end_line: int) -> str | None:
-    if not start_line or not end_line:
-        return None
-    try:
-        if isinstance(start_line, str):
-            start_line = int(start_line)
-        if isinstance(end_line, str):
-            end_line = int(end_line)
-        lines = file_lines(repo, file, max_tokens=None, use_local_files=True)
-        if lines:
-            lines = [""] + lines.splitlines()
-            return "\n".join(
-                lines[start_line: end_line + 1]
-            )
-    except Exception as e:
-        logging.error(
-            f"Error getting affected code block for {file} from {start_line} to {end_line}: {e}"
-        )
-    return None
-
-
-def provide_affected_code_blocks(issues: dict, repo: Repo):
-    for file, file_issues in issues.items():
-        for issue in file_issues:
-            for i in issue.get("affected_lines", []):
-                file_name = i.get("file", issue.get("file", file))
-                if block := get_affected_code_block(
-                    repo,
-                    file_name,
-                    i.get("start_line"),
-                    i.get("end_line")
-                ):
-                    i["affected_code"] = block
-
-
-async def review(
-    repo: Repo = None,
-    what: str = None,
-    against: str = None,
-    filters: str | list[str] = "",
-    use_merge_base: bool = True,
-    out_folder: str | os.PathLike | None = None,
-    pr: str | int = None
-):
-    try:
-        repo, cfg, diff, lines = _prepare(
-            repo=repo,
-            what=what,
-            against=against,
-            filters=filters,
-            use_merge_base=use_merge_base,
-            pr=pr,
-        )
-    except NoChangesInContextError:
-        logging.error("No changes to review")
-        return
-    responses = await mc.llm_parallel(
-        [
-            mc.prompt(
-                cfg.prompt,
-                input=file_diff,
-                file_lines=lines[file_diff.path],
-                **cfg.prompt_vars,
-            )
-            for file_diff in diff
-        ],
-        retries=cfg.retries,
-        parse_json=True,
-    )
-    issues = {file.path: issues for file, issues in zip(diff, responses) if issues}
-    provide_affected_code_blocks(issues, repo)
-    exec(cfg.post_process, {"mc": mc, **locals()})
-    out_folder = Path(out_folder or repo.working_tree_dir)
-    out_folder.mkdir(parents=True, exist_ok=True)
-    report = Report(issues=issues, number_of_processed_files=len(diff))
-    ctx = Context(
-        report=report,
-        config=cfg,
-        diff=diff,
-        repo=repo,
-    )
-    if cfg.pipeline_steps:
-        pipe = Pipeline(
-            ctx=ctx,
-            steps=cfg.pipeline_steps
-        )
-        pipe.run()
-    else:
-        logging.info("No pipeline steps defined, skipping pipeline execution")
-
-    report.summary = make_cr_summary(ctx)
-    report.save(file_name=out_folder / JSON_REPORT_FILE_NAME)
-    report_text = report.render(cfg, Report.Format.MARKDOWN)
-    text_report_path = out_folder / "code-review-report.md"
-    text_report_path.write_text(report_text, encoding="utf-8")
-    report.to_cli()
-
-
-def answer(
-    question: str,
-    repo: Repo = None,
-    what: str = None,
-    against: str = None,
-    filters: str | list[str] = "",
-    use_merge_base: bool = True,
-    use_pipeline: bool = True,
-    prompt_file: str = None,
-    pr: str | int = None,
-    aux_files: list[str] = None,
-) -> str | None:
-    try:
-        repo, config, diff, lines = _prepare(
-            repo=repo,
-            what=what,
-            against=against,
-            filters=filters,
-            use_merge_base=use_merge_base,
-            pr=pr
-        )
-    except NoChangesInContextError:
-        logging.error("No changes to review")
-        return
-
-    ctx = Context(
-        repo=repo,
-        diff=diff,
-        config=config,
-        report=Report()
-    )
-    if use_pipeline:
-        pipe = Pipeline(
-            ctx=ctx,
-            steps=config.pipeline_steps
-        )
-        pipe.run()
-
-    if aux_files or config.aux_files:
-        aux_files_dict = read_files(
-            repo,
-            (aux_files or list()) + config.aux_files,
-            config.max_code_tokens // 2
-        )
-    else:
-        aux_files_dict = dict()
-
-    if not prompt_file and config.answer_prompt.startswith("tpl:"):
-        prompt_file = str(config.answer_prompt)[4:]
-
-    if prompt_file:
-        prompt_func = partial(mc.tpl, prompt_file)
-    else:
-        prompt_func = partial(mc.prompt, config.answer_prompt)
-    prompt = prompt_func(
-        question=question,
-        diff=diff,
-        all_file_lines=lines,
-        pipeline_out=ctx.pipeline_out,
-        aux_files=aux_files_dict,
-        **config.prompt_vars,
-    )
-    response = mc.llm(
-        prompt,
-        callback=make_streaming_function() if Env.verbosity == 0 else None,
-    )
-    return response
+import os
+import fnmatch
+import logging
+from typing import Iterable
+from pathlib import Path
+from functools import partial
+
+import microcore as mc
+from gito.constants import REFS_VALUE_ALL
+from gito.gh_api import gh_api
+from microcore import ui
+from git import Repo, Commit
+from git.exc import GitCommandError
+from unidiff import PatchSet, PatchedFile
+from unidiff.constants import DEV_NULL
+
+from .context import Context
+from .project_config import ProjectConfig
+from .report_struct import Report
+from .constants import JSON_REPORT_FILE_NAME
+from .utils import make_streaming_function
+from .pipeline import Pipeline
+from .env import Env
+
+
+def review_subject_is_index(what):
+    return not what or what == 'INDEX'
+
+
+def is_binary_file(repo: Repo, file_path: str) -> bool:
+    """
+    Check if a file is binary by attempting to read it as text.
+    Returns True if the file is binary, False otherwise.
+    """
+    try:
+        # Attempt to read the file content from the repository tree
+        content = repo.tree()[file_path].data_stream.read()
+        # Try decoding as UTF-8; if it fails, it's likely binary
+        content.decode("utf-8")
+        return False
+    except KeyError:
+        try:
+            fs_path = Path(repo.working_tree_dir) / file_path
+            fs_path.read_text(encoding='utf-8')
+            return False
+        except FileNotFoundError:
+            logging.error(f"File {file_path} not found in the repository.")
+            return True
+        except UnicodeDecodeError:
+            return True
+        except Exception as e:
+            logging.error(f"Error reading file {file_path}: {e}")
+            return True
+    except UnicodeDecodeError:
+        return True
+    except Exception as e:
+        logging.warning(f"Error checking if file {file_path} is binary: {e}")
+        return True  # Conservatively treat errors as binary to avoid issues
+
+
+def commit_in_branch(repo: Repo, commit: Commit, target_branch: str) -> bool:
+    try:
+        # exit code 0 if commit is ancestor of branch
+        repo.git.merge_base('--is-ancestor', commit.hexsha, target_branch)
+        return True
+    except GitCommandError:
+        pass
+    return False
+
+
+def get_base_branch(repo: Repo, pr: int | str = None):
+    if os.getenv('GITHUB_ACTIONS'):
+
+        # triggered from PR
+        if base_ref := os.getenv('GITHUB_BASE_REF'):
+            logging.info(f"Using GITHUB_BASE_REF:{base_ref} as base branch")
+            return f'origin/{base_ref}'
+        logging.info("GITHUB_BASE_REF is not available...")
+        if pr:
+            api = gh_api(repo=repo)
+            pr_obj = api.pulls.get(pr)
+            logging.info(
+                f"Using 'origin/{pr_obj.base.ref}' as base branch "
+                f"(received via GH API for PR#{pr})"
+            )
+            return f'origin/{pr_obj.base.ref}'
+
+    try:
+        logging.info(
+            "Trying to resolve base branch from repo.remotes.origin.refs.HEAD.reference.name..."
+        )
+        # 'origin/main', 'origin/master', etc
+        # Stopped working in github actions since 07/2025
+        return repo.remotes.origin.refs.HEAD.reference.name
+    except AttributeError:
+        try:
+            logging.info(
+                "Checking if repo has 'main' or 'master' branchs to use as --against branch..."
+            )
+            remote_refs = repo.remotes.origin.refs
+            for branch_name in ['main', 'master']:
+                if hasattr(remote_refs, branch_name):
+                    return f'origin/{branch_name}'
+        except Exception:
+            pass
+
+        logging.error("Could not determine default branch from remote refs.")
+        raise ValueError("No default branch found in the repository.")
+
+
+def get_diff(
+    repo: Repo = None,
+    what: str = None,
+    against: str = None,
+    use_merge_base: bool = True,
+    pr: str | int = None
+) -> PatchSet | list[PatchedFile]:
+    repo = repo or Repo(".")
+    if what == REFS_VALUE_ALL:
+        what = get_base_branch(repo, pr=pr)
+        # Git's canonical empty tree hash
+        against = "4b825dc642cb6eb9a060e54bf8d69288fbee4904"
+        use_merge_base = False
+    if not against:
+        # 'origin/main', 'origin/master', etc
+        against = get_base_branch(repo, pr=pr)
+    if review_subject_is_index(what):
+        what = None  # working copy
+    if use_merge_base:
+        try:
+            if review_subject_is_index(what):
+                try:
+                    current_ref = repo.active_branch.name
+                except TypeError:
+                    # In detached HEAD state, use HEAD directly
+                    current_ref = "HEAD"
+                    logging.info(
+                        "Detected detached HEAD state, using HEAD as current reference"
+                    )
+            else:
+                current_ref = what
+            merge_base = repo.merge_base(current_ref or repo.active_branch.name, against)[0]
+            logging.info(
+                f"Merge base({ui.green(current_ref)},{ui.yellow(against)})"
+                f" --> {ui.cyan(merge_base.hexsha)}"
+            )
+            # if branch is already an ancestor of "against", merge_base == branch ⇒ it’s been merged
+            if merge_base.hexsha == repo.commit(current_ref or repo.active_branch.name).hexsha:
+                # @todo: check case: reviewing working copy index in main branch #103
+                logging.info(
+                    f"Branch is already merged. ({ui.green(current_ref)} vs {ui.yellow(against)})"
+                )
+                merge_sha = repo.git.log(
+                    '--merges',
+                    '--ancestry-path',
+                    f'{current_ref}..{against}',
+                    '-n',
+                    '1',
+                    '--pretty=format:%H'
+                ).strip()
+                if merge_sha:
+                    logging.info(f"Merge commit is {ui.cyan(merge_sha)}")
+                    merge_commit = repo.commit(merge_sha)
+
+                    other_merge_parent = None
+                    for parent in merge_commit.parents:
+                        logging.info(f"Checking merge parent: {parent.hexsha[:8]}")
+                        if parent.hexsha == merge_base.hexsha:
+                            logging.info(f"merge parent is {ui.cyan(parent.hexsha[:8])}, skipping")
+                            continue
+                        if not commit_in_branch(repo, parent, against):
+                            logging.warning(f"merge parent is not in {against}, skipping")
+                            continue
+                        logging.info(f"Found other merge parent: {ui.cyan(parent.hexsha[:8])}")
+                        other_merge_parent = parent
+                        break
+                    if other_merge_parent:
+                        first_common_ancestor = repo.merge_base(other_merge_parent, merge_base)[0]
+                        # for gito remote (feature_branch vs origin/main)
+                        # the same merge base appears in first_common_ancestor again
+                        if first_common_ancestor.hexsha == merge_base.hexsha:
+                            if merge_base.parents:
+                                first_common_ancestor = repo.merge_base(
+                                    other_merge_parent, merge_base.parents[0]
+                                )[0]
+                            else:
+                                logging.error(
+                                    "merge_base has no parents, "
+                                    "using merge_base as first_common_ancestor"
+                                )
+                        logging.info(
+                            f"{what} will be compared to "
+                            f"first common ancestor of {what} and {against}: "
+                            f"{ui.cyan(first_common_ancestor.hexsha[:8])}"
+                        )
+                        against = first_common_ancestor.hexsha
+                    else:
+                        logging.error(f"Can't find other merge parent for {merge_sha}")
+                else:
+                    logging.warning(
+                        f"No merge‐commit found for {current_ref!r}→{against!r}; "
+                        "falling back to merge‐base diff"
+                    )
+            else:
+                # normal case: branch not yet merged
+                against = merge_base.hexsha
+                logging.info(
+                    f"Using merge base: {ui.cyan(merge_base.hexsha[:8])} ({merge_base.summary})"
+                )
+        except Exception as e:
+            logging.error(f"Error finding merge base: {e}")
+    logging.info(
+        f"Making diff: {ui.green(what or 'INDEX')} vs {ui.yellow(against)}"
+    )
+    diff_content = repo.git.diff(against, what)
+    diff = PatchSet.from_string(diff_content)
+
+    # Filter out binary files
+    non_binary_diff = PatchSet([])
+    for patched_file in diff:
+        # Check if the file is binary using the source or target file path
+        file_path = (
+            patched_file.target_file
+            if patched_file.target_file != DEV_NULL
+            else patched_file.source_file
+        )
+        if file_path == DEV_NULL:
+            continue
+        if is_binary_file(repo, file_path.removeprefix("b/")):
+            logging.info(f"Skipping binary file: {patched_file.path}")
+            continue
+        non_binary_diff.append(patched_file)
+    return non_binary_diff
+
+
+def filter_diff(
+    patch_set: PatchSet | Iterable[PatchedFile], filters: str | list[str]
+) -> PatchSet | Iterable[PatchedFile]:
+    """
+    Filter the diff files by the given fnmatch filters.
+    """
+    assert isinstance(filters, (list, str))
+    if not isinstance(filters, list):
+        filters = [f.strip() for f in filters.split(",") if f.strip()]
+    if not filters:
+        return patch_set
+    files = [
+        file
+        for file in patch_set
+        if any(fnmatch.fnmatch(file.path, pattern) for pattern in filters)
+    ]
+    return files
+
+
+def read_file(repo: Repo, file: str, use_local_files: bool = False) -> str:
+    if use_local_files:
+        file_path = Path(repo.working_tree_dir) / file
+        try:
+            return file_path.read_text(encoding='utf-8')
+        except (FileNotFoundError, UnicodeDecodeError) as e:
+            logging.warning(f"Could not read file {file} from working directory: {e}")
+
+    # Read from HEAD (committed version)
+    return repo.tree()[file].data_stream.read().decode('utf-8')
+
+
+def file_lines(repo: Repo, file: str, max_tokens: int = None, use_local_files: bool = False) -> str:
+    text = read_file(repo=repo, file=file, use_local_files=use_local_files)
+    lines = [f"{i + 1}: {line}\n" for i, line in enumerate(text.splitlines())]
+    if max_tokens:
+        lines, removed_qty = mc.tokenizing.fit_to_token_size(lines, max_tokens)
+        if removed_qty:
+            lines.append(
+                f"(!) DISPLAYING ONLY FIRST {len(lines)} LINES DUE TO LARGE FILE SIZE\n"
+            )
+    return "".join(lines)
+
+
+def read_files(repo: Repo, files: list[str], max_tokens: int = None) -> dict:
+    out = dict()
+    total_tokens = 0
+    for file in files:
+        content = read_file(repo=repo, file=file, use_local_files=True)
+        total_tokens += mc.tokenizing.num_tokens_from_string(file)
+        total_tokens += mc.tokenizing.num_tokens_from_string(content)
+        if max_tokens and total_tokens > max_tokens:
+            logging.warning(
+                f"Skipping file {file} due to exceeding max_tokens limit ({max_tokens})"
+            )
+            continue
+        out[file] = content
+    return out
+
+
+def make_cr_summary(ctx: Context, **kwargs) -> str:
+    return (
+        mc.prompt(
+            ctx.config.summary_prompt,
+            diff=mc.tokenizing.fit_to_token_size(ctx.diff, ctx.config.max_code_tokens)[0],
+            issues=ctx.report.issues,
+            pipeline_out=ctx.pipeline_out,
+            env=Env,
+            **ctx.config.prompt_vars,
+            **kwargs,
+        ).to_llm()
+        if ctx.config.summary_prompt
+        else ""
+    )
+
+
+class NoChangesInContextError(Exception):
+    """
+    Exception raised when there are no changes in the context to review /answer questions.
+    """
+
+
+def _prepare(
+    repo: Repo = None,
+    what: str = None,
+    against: str = None,
+    filters: str | list[str] = "",
+    use_merge_base: bool = True,
+    pr: str | int = None,
+):
+    repo = repo or Repo(".")
+    cfg = ProjectConfig.load_for_repo(repo)
+    diff = get_diff(
+        repo=repo, what=what, against=against, use_merge_base=use_merge_base, pr=pr,
+    )
+    diff = filter_diff(diff, filters)
+    if not diff:
+        raise NoChangesInContextError()
+    lines = {
+        file_diff.path: (
+            file_lines(
+                repo,
+                file_diff.path,
+                cfg.max_code_tokens
+                - mc.tokenizing.num_tokens_from_string(str(file_diff)),
+                use_local_files=review_subject_is_index(what) or what == REFS_VALUE_ALL
+            )
+            if (
+                   file_diff.target_file != DEV_NULL and not file_diff.is_added_file
+               ) or what == REFS_VALUE_ALL
+            else ""
+        )
+        for file_diff in diff
+    }
+    return repo, cfg, diff, lines
+
+
+def get_affected_code_block(repo: Repo, file: str, start_line: int, end_line: int) -> str | None:
+    if not start_line or not end_line:
+        return None
+    try:
+        if isinstance(start_line, str):
+            start_line = int(start_line)
+        if isinstance(end_line, str):
+            end_line = int(end_line)
+        lines = file_lines(repo, file, max_tokens=None, use_local_files=True)
+        if lines:
+            lines = [""] + lines.splitlines()
+            return "\n".join(
+                lines[start_line: end_line + 1]
+            )
+    except Exception as e:
+        logging.error(
+            f"Error getting affected code block for {file} from {start_line} to {end_line}: {e}"
+        )
+    return None
+
+
+def provide_affected_code_blocks(issues: dict, repo: Repo):
+    for file, file_issues in issues.items():
+        for issue in file_issues:
+            for i in issue.get("affected_lines", []):
+                file_name = i.get("file", issue.get("file", file))
+                if block := get_affected_code_block(
+                    repo,
+                    file_name,
+                    i.get("start_line"),
+                    i.get("end_line")
+                ):
+                    i["affected_code"] = block
+
+
+async def review(
+    repo: Repo = None,
+    what: str = None,
+    against: str = None,
+    filters: str | list[str] = "",
+    use_merge_base: bool = True,
+    out_folder: str | os.PathLike | None = None,
+    pr: str | int = None
+):
+    reviewing_all = what == REFS_VALUE_ALL
+    try:
+        repo, cfg, diff, lines = _prepare(
+            repo=repo,
+            what=what,
+            against=against,
+            filters=filters,
+            use_merge_base=use_merge_base,
+            pr=pr,
+        )
+    except NoChangesInContextError:
+        logging.error("No changes to review")
+        return
+    responses = await mc.llm_parallel(
+        [
+            mc.prompt(
+                cfg.prompt,
+                input=(
+                    file_diff if not reviewing_all
+                    else str(file_diff.path) + ":\n" + lines[file_diff.path]
+                ),
+                file_lines=lines[file_diff.path] if not reviewing_all else None,
+                **cfg.prompt_vars,
+            )
+            for file_diff in diff
+        ],
+        retries=cfg.retries,
+        parse_json=True,
+    )
+    issues = {file.path: issues for file, issues in zip(diff, responses) if issues}
+    provide_affected_code_blocks(issues, repo)
+    exec(cfg.post_process, {"mc": mc, **locals()})
+    out_folder = Path(out_folder or repo.working_tree_dir)
+    out_folder.mkdir(parents=True, exist_ok=True)
+    report = Report(issues=issues, number_of_processed_files=len(diff))
+    ctx = Context(
+        report=report,
+        config=cfg,
+        diff=diff,
+        repo=repo,
+    )
+    if cfg.pipeline_steps:
+        pipe = Pipeline(
+            ctx=ctx,
+            steps=cfg.pipeline_steps
+        )
+        pipe.run()
+    else:
+        logging.info("No pipeline steps defined, skipping pipeline execution")
+
+    report.summary = make_cr_summary(ctx)
+    report.save(file_name=out_folder / JSON_REPORT_FILE_NAME)
+    report_text = report.render(cfg, Report.Format.MARKDOWN)
+    text_report_path = out_folder / "code-review-report.md"
+    text_report_path.write_text(report_text, encoding="utf-8")
+    report.to_cli()
+
+
+def answer(
+    question: str,
+    repo: Repo = None,
+    what: str = None,
+    against: str = None,
+    filters: str | list[str] = "",
+    use_merge_base: bool = True,
+    use_pipeline: bool = True,
+    prompt_file: str = None,
+    pr: str | int = None,
+    aux_files: list[str] = None,
+) -> str | None:
+    try:
+        repo, config, diff, lines = _prepare(
+            repo=repo,
+            what=what,
+            against=against,
+            filters=filters,
+            use_merge_base=use_merge_base,
+            pr=pr
+        )
+    except NoChangesInContextError:
+        logging.error("No changes to review")
+        return
+
+    ctx = Context(
+        repo=repo,
+        diff=diff,
+        config=config,
+        report=Report()
+    )
+    if use_pipeline:
+        pipe = Pipeline(
+            ctx=ctx,
+            steps=config.pipeline_steps
+        )
+        pipe.run()
+
+    if aux_files or config.aux_files:
+        aux_files_dict = read_files(
+            repo,
+            (aux_files or list()) + config.aux_files,
+            config.max_code_tokens // 2
+        )
+    else:
+        aux_files_dict = dict()
+
+    if not prompt_file and config.answer_prompt.startswith("tpl:"):
+        prompt_file = str(config.answer_prompt)[4:]
+
+    if prompt_file:
+        prompt_func = partial(mc.tpl, prompt_file)
+    else:
+        prompt_func = partial(mc.prompt, config.answer_prompt)
+    prompt = prompt_func(
+        question=question,
+        diff=diff,
+        all_file_lines=lines,
+        pipeline_out=ctx.pipeline_out,
+        aux_files=aux_files_dict,
+        **config.prompt_vars,
+    )
+    response = mc.llm(
+        prompt,
+        callback=make_streaming_function() if Env.verbosity == 0 else None,
+    )
+    return response