xiaxinhang/pr-agent
1
0
Fork 0
mirror of https://github.com/qodo-ai/pr-agent.git synced 2025-07-02 20:00:41 +08:00
Code Issues Packages Projects Releases Wiki Activity

Merge commit 'e4f177908b620e46740b03966fda9243473d979e' into hl/pr_review_table

Browse Source
This commit is contained in:
Hussam.lawen
2024-02-08 14:26:29 +02:00
parent ddb89a7474 e4f177908b
commit c2088b7752
41 changed files with 653 additions and 383 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
pr_agent/algo/__init__.py
View File

@ -9,6 +9,7 @@ MAX_TOKENS = {
'gpt-4-0613': 8000,
'gpt-4-32k': 32000,
'gpt-4-1106-preview': 128000, # 128K, but may be limited by config.max_model_tokens
'gpt-4-0125-preview': 128000, # 128K, but may be limited by config.max_model_tokens
'claude-instant-1': 100000,
'claude-2': 100000,
'command-nightly': 4096,

12
pr_agent/algo/git_patch_processing.py
View File

@ -3,7 +3,7 @@ from __future__ import annotations
import re
from pr_agent.config_loader import get_settings
from pr_agent.git_providers.git_provider import EDIT_TYPE
from pr_agent.algo.types import EDIT_TYPE, FilePatchInfo
from pr_agent.log import get_logger
@ -181,7 +181,7 @@ __old hunk__
...
"""
patch_with_lines_str = f"\n\n## {file.filename}\n"
patch_with_lines_str = f"\n\n## file: '{file.filename.strip()}'\n"
patch_lines = patch.splitlines()
RE_HUNK_HEADER = re.compile(
r"^@@ -(\d+)(?:,(\d+))? \+(\d+)(?:,(\d+))? @@[ ]?(.*)")
@ -202,11 +202,11 @@ __old hunk__
if new_content_lines:
if prev_header_line:
patch_with_lines_str += f'\n{prev_header_line}\n'
patch_with_lines_str += '__new hunk__\n'
patch_with_lines_str = patch_with_lines_str.rstrip()+'\n__new hunk__\n'
for i, line_new in enumerate(new_content_lines):
patch_with_lines_str += f"{start2 + i} {line_new}\n"
if old_content_lines:
patch_with_lines_str += '__old hunk__\n'
patch_with_lines_str = patch_with_lines_str.rstrip()+'\n__old hunk__\n'
for line_old in old_content_lines:
patch_with_lines_str += f"{line_old}\n"
new_content_lines = []
@ -236,11 +236,11 @@ __old hunk__
if match and new_content_lines:
if new_content_lines:
patch_with_lines_str += f'\n{header_line}\n'
patch_with_lines_str += '\n__new hunk__\n'
patch_with_lines_str = patch_with_lines_str.rstrip()+ '\n__new hunk__\n'
for i, line_new in enumerate(new_content_lines):
patch_with_lines_str += f"{start2 + i} {line_new}\n"
if old_content_lines:
patch_with_lines_str += '\n__old hunk__\n'
patch_with_lines_str = patch_with_lines_str.rstrip() + '\n__old hunk__\n'
for line_old in old_content_lines:
patch_with_lines_str += f"{line_old}\n"

108
pr_agent/algo/pr_processing.py
View File

@ -1,9 +1,7 @@
from __future__ import annotations
import difflib
import re
import traceback
from typing import Any, Callable, List, Tuple
from typing import Callable, List, Tuple
from github import RateLimitExceededException
@ -11,9 +9,10 @@ from pr_agent.algo.git_patch_processing import convert_to_hunks_with_lines_numbe
from pr_agent.algo.language_handler import sort_files_by_main_languages
from pr_agent.algo.file_filter import filter_ignored
from pr_agent.algo.token_handler import TokenHandler
from pr_agent.algo.utils import get_max_tokens
from pr_agent.algo.utils import get_max_tokens, ModelType
from pr_agent.config_loader import get_settings
from pr_agent.git_providers.git_provider import FilePatchInfo, GitProvider, EDIT_TYPE
from pr_agent.git_providers.git_provider import GitProvider
from pr_agent.algo.types import EDIT_TYPE, FilePatchInfo
from pr_agent.log import get_logger
DELETED_FILES_ = "Deleted files:\n"
@ -209,9 +208,9 @@ def pr_generate_compressed_diff(top_langs: list, token_handler: TokenHandler, mo
if patch:
if not convert_hunks_to_line_numbers:
patch_final = f"## {file.filename}\n\n{patch}\n"
patch_final = f"\n\n## file: '{file.filename.strip()}\n\n{patch.strip()}\n'"
else:
patch_final = patch
patch_final = "\n\n" + patch.strip()
patches.append(patch_final)
total_tokens += token_handler.count_tokens(patch_final)
if get_settings().config.verbosity_level >= 2:
@ -220,8 +219,8 @@ def pr_generate_compressed_diff(top_langs: list, token_handler: TokenHandler, mo
return patches, modified_files_list, deleted_files_list, added_files_list
async def retry_with_fallback_models(f: Callable):
all_models = _get_all_models()
async def retry_with_fallback_models(f: Callable, model_type: ModelType = ModelType.REGULAR):
all_models = _get_all_models(model_type)
all_deployments = _get_all_deployments(all_models)
# try each (model, deployment_id) pair until one is successful, otherwise raise exception
for i, (model, deployment_id) in enumerate(zip(all_models, all_deployments)):
@ -243,8 +242,11 @@ async def retry_with_fallback_models(f: Callable):
raise # Re-raise the last exception
def _get_all_models() -> List[str]:
model = get_settings().config.model
def _get_all_models(model_type: ModelType = ModelType.REGULAR) -> List[str]:
if model_type == ModelType.TURBO:
model = get_settings().config.model_turbo
else:
model = get_settings().config.model
fallback_models = get_settings().config.fallback_models
if not isinstance(fallback_models, list):
fallback_models = [m.strip() for m in fallback_models.split(",")]
@ -267,78 +269,6 @@ def _get_all_deployments(all_models: List[str]) -> List[str]:
return all_deployments
def find_line_number_of_relevant_line_in_file(diff_files: List[FilePatchInfo],
relevant_file: str,
relevant_line_in_file: str,
absolute_position: int = None) -> Tuple[int, int]:
position = -1
if absolute_position is None:
absolute_position = -1
re_hunk_header = re.compile(
r"^@@ -(\d+)(?:,(\d+))? \+(\d+)(?:,(\d+))? @@[ ]?(.*)")
for file in diff_files:
if file.filename and (file.filename.strip() == relevant_file):
patch = file.patch
patch_lines = patch.splitlines()
delta = 0
start1, size1, start2, size2 = 0, 0, 0, 0
if absolute_position != -1: # matching absolute to relative
for i, line in enumerate(patch_lines):
# new hunk
if line.startswith('@@'):
delta = 0
match = re_hunk_header.match(line)
start1, size1, start2, size2 = map(int, match.groups()[:4])
elif not line.startswith('-'):
delta += 1
#
absolute_position_curr = start2 + delta - 1
if absolute_position_curr == absolute_position:
position = i
break
else:
# try to find the line in the patch using difflib, with some margin of error
matches_difflib: list[str | Any] = difflib.get_close_matches(relevant_line_in_file,
patch_lines, n=3, cutoff=0.93)
if len(matches_difflib) == 1 and matches_difflib[0].startswith('+'):
relevant_line_in_file = matches_difflib[0]
for i, line in enumerate(patch_lines):
if line.startswith('@@'):
delta = 0
match = re_hunk_header.match(line)
start1, size1, start2, size2 = map(int, match.groups()[:4])
elif not line.startswith('-'):
delta += 1
if relevant_line_in_file in line and line[0] != '-':
position = i
absolute_position = start2 + delta - 1
break
if position == -1 and relevant_line_in_file[0] == '+':
no_plus_line = relevant_line_in_file[1:].lstrip()
for i, line in enumerate(patch_lines):
if line.startswith('@@'):
delta = 0
match = re_hunk_header.match(line)
start1, size1, start2, size2 = map(int, match.groups()[:4])
elif not line.startswith('-'):
delta += 1
if no_plus_line in line and line[0] != '-':
# The model might add a '+' to the beginning of the relevant_line_in_file even if originally
# it's a context line
position = i
absolute_position = start2 + delta - 1
break
return position, absolute_position
def get_pr_multi_diffs(git_provider: GitProvider,
token_handler: TokenHandler,
model: str,
@ -375,6 +305,13 @@ def get_pr_multi_diffs(git_provider: GitProvider,
for lang in pr_languages:
sorted_files.extend(sorted(lang['files'], key=lambda x: x.tokens, reverse=True))
# try first a single run with standard diff string, with patch extension, and no deletions
patches_extended, total_tokens, patches_extended_tokens = pr_generate_extended_diff(
pr_languages, token_handler, add_line_numbers_to_hunks=True)
if total_tokens + OUTPUT_BUFFER_TOKENS_SOFT_THRESHOLD < get_max_tokens(model):
return ["\n".join(patches_extended)]
patches = []
final_diff_list = []
total_tokens = token_handler.prompt_tokens
@ -398,6 +335,11 @@ def get_pr_multi_diffs(git_provider: GitProvider,
patch = convert_to_hunks_with_lines_numbers(patch, file)
new_patch_tokens = token_handler.count_tokens(patch)
if patch and (token_handler.prompt_tokens + new_patch_tokens) > get_max_tokens(model) - OUTPUT_BUFFER_TOKENS_SOFT_THRESHOLD:
get_logger().warning(f"Patch too large, skipping: {file.filename}")
continue
if patch and (total_tokens + new_patch_tokens > get_max_tokens(model) - OUTPUT_BUFFER_TOKENS_SOFT_THRESHOLD):
final_diff = "\n".join(patches)
final_diff_list.append(final_diff)

23
pr_agent/algo/types.py Normal file
View File

@ -0,0 +1,23 @@
from dataclasses import dataclass
from enum import Enum
class EDIT_TYPE(Enum):
ADDED = 1
DELETED = 2
MODIFIED = 3
RENAMED = 4
UNKNOWN = 5
@dataclass
class FilePatchInfo:
base_file: str
head_file: str
patch: str
filename: str
tokens: int = -1
edit_type: EDIT_TYPE = EDIT_TYPE.UNKNOWN
old_filename: str = None
num_plus_lines: int = -1
num_minus_lines: int = -1

81
pr_agent/algo/utils.py
View File

@ -5,7 +5,8 @@ import json
import re
import textwrap
from datetime import datetime
from typing import Any, List
from enum import Enum
from typing import Any, List, Tuple
import yaml
from starlette_context import context
@ -13,8 +14,12 @@ from starlette_context import context
from pr_agent.algo import MAX_TOKENS
from pr_agent.algo.token_handler import get_token_encoder
from pr_agent.config_loader import get_settings, global_settings
from pr_agent.algo.types import FilePatchInfo
from pr_agent.log import get_logger
class ModelType(str, Enum):
REGULAR = "regular"
TURBO = "turbo"
def get_setting(key: str) -> Any:
try:
@ -489,4 +494,76 @@ def replace_code_tags(text):
parts = text.split('`')
for i in range(1, len(parts), 2):
parts[i] = '<code>' + parts[i] + '</code>'
return ''.join(parts)
return ''.join(parts)
def find_line_number_of_relevant_line_in_file(diff_files: List[FilePatchInfo],
relevant_file: str,
relevant_line_in_file: str,
absolute_position: int = None) -> Tuple[int, int]:
position = -1
if absolute_position is None:
absolute_position = -1
re_hunk_header = re.compile(
r"^@@ -(\d+)(?:,(\d+))? \+(\d+)(?:,(\d+))? @@[ ]?(.*)")
for file in diff_files:
if file.filename and (file.filename.strip() == relevant_file):
patch = file.patch
patch_lines = patch.splitlines()
delta = 0
start1, size1, start2, size2 = 0, 0, 0, 0
if absolute_position != -1: # matching absolute to relative
for i, line in enumerate(patch_lines):
# new hunk
if line.startswith('@@'):
delta = 0
match = re_hunk_header.match(line)
start1, size1, start2, size2 = map(int, match.groups()[:4])
elif not line.startswith('-'):
delta += 1
#
absolute_position_curr = start2 + delta - 1
if absolute_position_curr == absolute_position:
position = i
break
else:
# try to find the line in the patch using difflib, with some margin of error
matches_difflib: list[str | Any] = difflib.get_close_matches(relevant_line_in_file,
patch_lines, n=3, cutoff=0.93)
if len(matches_difflib) == 1 and matches_difflib[0].startswith('+'):
relevant_line_in_file = matches_difflib[0]
for i, line in enumerate(patch_lines):
if line.startswith('@@'):
delta = 0
match = re_hunk_header.match(line)
start1, size1, start2, size2 = map(int, match.groups()[:4])
elif not line.startswith('-'):
delta += 1
if relevant_line_in_file in line and line[0] != '-':
position = i
absolute_position = start2 + delta - 1
break
if position == -1 and relevant_line_in_file[0] == '+':
no_plus_line = relevant_line_in_file[1:].lstrip()
for i, line in enumerate(patch_lines):
if line.startswith('@@'):
delta = 0
match = re_hunk_header.match(line)
start1, size1, start2, size2 = map(int, match.groups()[:4])
elif not line.startswith('-'):
delta += 1
if no_plus_line in line and line[0] != '-':
# The model might add a '+' to the beginning of the relevant_line_in_file even if originally
# it's a context line
position = i
absolute_position = start2 + delta - 1
break
return position, absolute_position