refactor: code clean up

Author: williamfzc

Dockerfile

@@ -12,8 +12,7 @@ RUN apk add --no-cache python3 py3-pip && \
 # scip converter
 RUN curl -fLo scip-linux-amd64.tar.gz https://github.com/sourcegraph/scip/releases/download/v0.2.3/scip-linux-amd64.tar.gz \
   && tar xf ./scip-linux-amd64.tar.gz \
-  && chmod +x ./scip \
-  && ./scip --help
+  && chmod +x ./scip
 
 # java/kotlin/scala scip/lsif
 RUN apk add openjdk17 \

ai.py

@@ -0,0 +1,23 @@
+import openai
+
+
+def process_with_ai(raw: str) -> str:
+    req = f"""
+You are a bot for helping code review.
+Standard csv report was generated from diff analysis tool.
+
+Here is a csv report below for a specific commit.
+Evaluate it and indicate the most important parts which reviewers should care.
+Empty report means that there are no dangerous changes.
+
+--- report start ---
+
+{raw}
+
+--- report end ---
+"""
+
+    completion = openai.ChatCompletion.create(
+        model="gpt-3.5-turbo", messages=[{"role": "user", "content": req}]
+    )
+    return completion.choices[0].message.content

config.py

@@ -0,0 +1,5 @@
+user_dir = "/github/workspace"
+support_langs = {"golang", "python", "java", "kotlin"}
+csv_result_file = "./output.csv"
+json_result_file = "./output.json"
+dot_result_file = "./output.dot"

diff.py

@@ -0,0 +1,29 @@
+from config import user_dir, csv_result_file, json_result_file, dot_result_file
+from utils import check_call
+
+
+def gen_diff(before_sha: str, after_sha: str):
+    # gen diff
+    set_safe_git_dir()
+    check_call(
+        [
+            "srctx",
+            "diff",
+            "--src",
+            user_dir,
+            "--before",
+            before_sha,
+            "--after",
+            after_sha,
+            "--outputCsv",
+            csv_result_file,
+            "--outputJson",
+            json_result_file,
+            "--outputDot",
+            dot_result_file,
+        ]
+    )
+
+
+def set_safe_git_dir():
+    check_call(["git", "config", "--global", "--add", "safe.directory", user_dir])

index.py

@@ -0,0 +1,36 @@
+import os
+
+from loguru import logger
+
+from diff import set_safe_git_dir
+from config import user_dir
+from utils import check_call
+
+
+def gen_index(lang: str):
+    set_safe_git_dir()
+    files = os.listdir(user_dir)
+    logger.info(f"files: {files}")
+
+    if lang == "golang":
+        gen_golang_index()
+    elif lang == "python":
+        gen_py_index()
+    else:
+        logger.error("no index mapping")
+        return
+
+
+def gen_golang_index():
+    check_call(["lsif-go", "-v"])
+
+
+def gen_java_and_kotlin_index():
+    # https://sourcegraph.github.io/scip-java/docs/getting-started.html#run-scip-java-index
+    check_call(["scip-java", "index", "--output", "index.scip"])
+    # https://github.com/sourcegraph/scip/blob/main/docs/CLI.md
+    check_call(["scip", "convert", "--from", "index.scip", "--to", "dump.lsif"])
+
+
+def gen_py_index():
+    check_call(["lsif-py", ".", "--file", "./dump.lsif"])

main.py

@@ -5,93 +5,88 @@
 import sys
 
 import openai
-import requests
 from csvtomd import csv_to_table, md_table
 from loguru import logger
 
+from ai import process_with_ai
 from comment import send_comment
+from config import support_langs, csv_result_file, json_result_file, dot_result_file
+from diff import gen_diff
+from index import gen_index
 from object import FileList
 
-user_dir = "/github/workspace"
-support_langs = {"golang", "python"}
-csv_result_file = "./output.csv"
-json_result_file = "./output.json"
-dot_result_file = "./output.dot"
 
+def main():
+    args = sys.argv[1:]
+    lang = args[0]
+    before_sha = args[1]
+    after_sha = args[2]
+    repo_token = args[3]
+    issue_number = args[4]
+    openai_api_key = args[5]
 
-def gen_index(lang: str):
-    set_safe_git_dir()
-    files = os.listdir(user_dir)
-    logger.info(f"files: {files}")
-
-    if lang == "golang":
-        gen_golang_index()
-    elif lang == "python":
-        gen_py_index()
-    else:
-        logger.error("no index mapping")
+    # check
+    if lang not in support_langs:
+        logger.error(f"lang not supported: {lang}, supported: {support_langs}")
+        return
+    if not before_sha or not after_sha:
+        logger.warning("sha empty. Use default.")
+        before_sha = "HEAD~1"
+        after_sha = "HEAD"
+    if not repo_token:
+        logger.error("no repo token provided")
         return
 
+    # data prepare
+    gen_index(lang)
+    gen_diff(before_sha, after_sha)
 
-def gen_golang_index():
-    _check_call(["lsif-go", "-v"])
-
-
-def gen_py_index():
-    _check_call(["lsif-py", ".", "--file", "./dump.lsif"])
-
-
-def _check_call(commands: list):
-    subprocess.check_call(commands, cwd=user_dir)
-
-
-def set_safe_git_dir():
-    _check_call(["git", "config", "--global", "--add", "safe.directory", user_dir])
+    with open(csv_result_file, encoding="utf-8") as f:
+        content = f.read()
 
+    ai_content = "-"
+    if not openai_api_key:
+        logger.warning("no openai api key found. Use raw data.")
+    else:
+        logger.info("process with openai")
+        openai.api_key = openai_api_key
+        ai_content = process_with_ai(content)
 
-def gen_diff(before_sha: str, after_sha: str):
-    # gen diff
-    set_safe_git_dir()
-    _check_call(
-        [
-            "srctx",
-            "diff",
-            "--src",
-            user_dir,
-            "--before",
-            before_sha,
-            "--after",
-            after_sha,
-            "--outputCsv",
-            csv_result_file,
-            "--outputJson",
-            json_result_file,
-            "--outputDot",
-            dot_result_file,
-        ]
-    )
+    # todo: add to feedback?
+    logger.info(f"ai resp: {ai_content}")
 
+    repo_name = os.getenv("GITHUB_REPOSITORY")
+    process_json(json_result_file, csv_result_file)
+    diff_desc = f"Start from {before_sha} to {after_sha}."
+    summary = get_summary(json_result_file)
+    md_table_raw = convert_csv_to_md(csv_result_file)
 
-def process_with_ai(raw: str) -> str:
-    req = f"""
-You are a bot for helping code review.
-Standard csv report was generated from diff analysis tool.
+    # graph
+    svg_bytes = dot_to_svg(dot_result_file)
+    # todo: have no idea about how to display this graph in comment without any extra servers
 
-Here is a csv report below for a specific commit.
-Evaluate it and indicate the most important parts which reviewers should care.
-Empty report means that there are no dangerous changes.
+    final_content = f"""
+## [DiffCtx](https://github.com/williamfzc/diffctx) Report
 
---- report start ---
+{diff_desc}
 
-{raw}
+{summary}
 
---- report end ---
+{md_table_raw}
 """
+    logger.info(f"final comment: {final_content}")
 
-    completion = openai.ChatCompletion.create(
-        model="gpt-3.5-turbo", messages=[{"role": "user", "content": req}]
-    )
-    return completion.choices[0].message.content
+    # feedback
+    if not issue_number:
+        logger.warning(
+            "This action is not triggered by a PR. Will not leave any comments."
+        )
+        return
+    send_comment(repo_token, repo_name, int(issue_number), final_content)
+
+    # # code comments
+    # lines = parse_file_as(typing.List[LineStat], pathlib.Path(json_result_file))
+    # send_code_comments(repo_token, after_sha, repo_name, int(issue_number), lines)
 
 
 def convert_csv_to_md(csv_file) -> str:
@@ -112,9 +107,15 @@ def format_percentage(numerator, denominator):
         return f"{percent:.2f}% ({numerator}/{denominator})"
 
     for file in file_list.files:
-        file.affectedLinePercentRepr = format_percentage(file.affectedLines, file.totalLines)
-        file.affectedFunctionPercentRepr = format_percentage(file.affectedFunctions, file.totalFunctions)
-        file.affectedReferencePercentRepr = format_percentage(file.affectedReferences, file.totalReferences)
+        file.affectedLinePercentRepr = format_percentage(
+            file.affectedLines, file.totalLines
+        )
+        file.affectedFunctionPercentRepr = format_percentage(
+            file.affectedFunctions, file.totalFunctions
+        )
+        file.affectedReferencePercentRepr = format_percentage(
+            file.affectedReferences, file.totalReferences
+        )
 
     file_list.files = sort_files_by_impact(file_list.files)
 
@@ -124,7 +125,7 @@ def format_percentage(numerator, denominator):
         "AffectedFunctions",
         "AffectedReferences",
     ]
-    with open(output_csv, "w", newline='') as f:
+    with open(output_csv, "w", newline="") as f:
         writer = csv.DictWriter(f, fieldnames=cols)
         writer.writeheader()
 
@@ -149,10 +150,12 @@ def get_summary(input_json) -> str:
     affected_functions = sum([each.affectedFunctions for each in file_list.files])
     affected_refs = sum([each.affectedReferences for each in file_list.files])
 
-    return f"This commit directly influences {affected_files} files, " \
-           f"{affected_lines} lines, " \
-           f"{affected_functions} functions. " \
-           f"Indirectly influences {affected_refs} functions. "
+    return (
+        f"This commit directly influences {affected_files} files, "
+        f"{affected_lines} lines, "
+        f"{affected_functions} functions. "
+        f"Indirectly influences {affected_refs} functions. "
+    )
 
 
 def dot_to_svg(dot_file):
@@ -167,78 +170,5 @@ def sort_key(f):
     return sorted(files, key=sort_key, reverse=True)
 
 
-def main():
-    args = sys.argv[1:]
-    lang = args[0]
-    before_sha = args[1]
-    after_sha = args[2]
-    repo_token = args[3]
-    issue_number = args[4]
-    openai_api_key = args[5]
-
-    # check
-    if lang not in support_langs:
-        logger.error(f"lang not supported: {lang}, supported: {support_langs}")
-        return
-    if not before_sha or not after_sha:
-        logger.warning("sha empty. Use default.")
-        before_sha = "HEAD~1"
-        after_sha = "HEAD"
-    if not repo_token:
-        logger.error("no repo token provided")
-        return
-
-    # data prepare
-    gen_index(lang)
-    gen_diff(before_sha, after_sha)
-
-    with open(csv_result_file, encoding="utf-8") as f:
-        content = f.read()
-
-    ai_content = "-"
-    if not openai_api_key:
-        logger.warning("no openai api key found. Use raw data.")
-    else:
-        logger.info("process with openai")
-        openai.api_key = openai_api_key
-        ai_content = process_with_ai(content)
-
-    # todo: add to feedback?
-    logger.info(f"ai resp: {ai_content}")
-
-    repo_name = os.getenv("GITHUB_REPOSITORY")
-    process_json(json_result_file, csv_result_file)
-    diff_desc = f"Start from {before_sha} to {after_sha}."
-    summary = get_summary(json_result_file)
-    md_table_raw = convert_csv_to_md(csv_result_file)
-
-    # graph
-    svg_bytes = dot_to_svg(dot_result_file)
-    # todo: have no idea about how to display this graph in comment without any extra servers
-
-    final_content = f"""
-## [DiffCtx](https://github.com/williamfzc/diffctx) Report
-
-{diff_desc}
-
-{summary}
-
-{md_table_raw}
-"""
-    logger.info(f"final comment: {final_content}")
-
-    # feedback
-    if not issue_number:
-        logger.warning(
-            "This action is not triggered by a PR. Will not leave any comments."
-        )
-        return
-    send_comment(repo_token, repo_name, int(issue_number), final_content)
-
-    # # code comments
-    # lines = parse_file_as(typing.List[LineStat], pathlib.Path(json_result_file))
-    # send_code_comments(repo_token, after_sha, repo_name, int(issue_number), lines)
-
-
 if __name__ == "__main__":
     main()

utils.py

@@ -0,0 +1,7 @@
+import subprocess
+
+from config import user_dir
+
+
+def check_call(commands: list):
+    subprocess.check_call(commands, cwd=user_dir)