fix: regression ci (cloudwego#96)

* docs: README

* fix: python install

* fix: run first testdata in script

* tests: check run success in run_testdata

* tests: update regression to use auto env setup and fix exit value

* tests: add typescript to regression

* fix: ts ci

Author: Hoblovski

.github/workflows/regression.yml

@@ -12,6 +12,8 @@ jobs:
   run_all_tests:
     runs-on: ubuntu-latest
     #if: "!contains(github.event.pull_request.title, '[NO-REGRESSION-TEST]')"
+    env:
+      LANGS: "go rust python java typescript"
     steps:
       - name: Setup Go environment
         uses: actions/setup-go@v5
@@ -29,16 +31,21 @@ jobs:
         with:
           python-version: '3.11'
 
+      - name: Setup JDK 21
+        uses: actions/setup-java@v4
+        with:
+          java-version: '21'
+          distribution: 'temurin'
+
+      - name: Setup Node.js
+        uses: actions/setup-node@v4
+        with:
+          node-version: '22'
+
       - name: Checkout pull request code
         uses: actions/checkout@v4
         with:
           path: 'pr_repo'
-          submodules: true
-
-      - name: Install Python dependencies
-        run: |
-          pip install -r ./pr_repo/script/requirements.txt
-          pip install ./pr_repo/pylsp
 
       - name: Checkout main branch code
         uses: actions/checkout@v4
@@ -51,14 +58,26 @@ jobs:
           (cd main_repo && go build -o ../abcoder_old)
           (cd pr_repo && go build -o ../abcoder_new)
 
-      - name: Run test scripts and generate outputs
+      - name: Install evaluation dependencies
+        run: pip install -r ./pr_repo/script/requirements.txt
+
+      - name: Install LSPs
         run: |
-          LANGS="go rust python" OUTDIR=out_old ABCEXE=./abcoder_old ./pr_repo/script/run_all_testdata.sh
-          LANGS="go rust python" OUTDIR=out_new ABCEXE=./abcoder_new ./pr_repo/script/run_all_testdata.sh
+          OUTDIR=out_new ABCEXE=./abcoder_new ./pr_repo/script/run_testdata.sh first
+          # use the same JDTLS for consistency and to avoid wasting time installing a duplicate JDTLS
+          echo "JDTLS_ROOT_PATH=$(realpath ./pr_repo/lang/java/lsp/jdtls/jdt-language-server-*)" >> $GITHUB_ENV
+
+      - name: Run OLD abcoder
+        run:
+          OUTDIR=out_old ABCEXE=./abcoder_old ./pr_repo/script/run_testdata.sh all
+
+      - name: Run NEW abcoder
+        run:
+          OUTDIR=out_new ABCEXE=./abcoder_new ./pr_repo/script/run_testdata.sh all
 
       - name: Compare outputs and check for regression
         id: diff_check
-        run: ./pr_repo/script/diffjson.py out_old out_new
+        run: ./pr_repo/script/diffjson.py out_old out_new || echo "failed=true" >> $GITHUB_OUTPUT
         continue-on-error: true
 
       - name: Upload output directories
@@ -70,3 +89,7 @@ jobs:
             out_old
             out_new
           retention-days: 3
+
+      - name: Status check
+        if: steps.diff_check.outputs.failed == 'true'
+        run: exit 1

README.md

@@ -36,17 +36,19 @@ see [UniAST Specification](docs/uniast-zh.md)
 2. Use ABCoder to parse a repository to UniAST (JSON)
 
     ```bash
-    abcoder parse {language} {repo-path} > xxx.json
+    abcoder parse {language} {repo-path} -o xxx.json
     ```
 
+    ABCoder will try to install any dependency automatically.
+    In case of failure (or if you want to customize installation), refer to the [docs](./docs/lsp-installation-en.md).
+
     For example, to parse a Go repository:
 
     ```bash
     git clone https://github.com/cloudwego/localsession.git localsession
     abcoder parse go localsession -o /abcoder-asts/localsession.json
     ```
 
-    To parse repositories in other languages, [install the corresponding language server first](./docs/lsp-installation-en.md).
 
 3. Integrate ABCoder's MCP tools into your AI agent.
 

lang/python/lib.go

@@ -16,7 +16,6 @@ package python
 
 import (
 	"fmt"
-	"os"
 	"os/exec"
 	"regexp"
 	"strconv"
@@ -63,10 +62,6 @@ func InstallLanguageServer() (string, error) {
 		log.Info("pylsp already installed: %v", out)
 		return lspName, nil
 	}
-	if _, err := os.Stat("go.mod"); os.IsNotExist(err) {
-		log.Error("Auto installation requires working directory to be /path/to/abcoder/")
-		return "", fmt.Errorf("bad cwd")
-	}
 	if err := CheckPythonVersion(); err != nil {
 		log.Error("python version check failed: %v", err)
 		return "", err

script/run_all_testdata.sh

@@ -0,0 +1,81 @@
+#!/bin/bash
+# Generate uniast for all testdata.
+#
+# USAGE:
+# 1. Save the uniast for all testdata to out/
+# $ OUTDIR=out/ ./script/run_testdata.sh all
+#
+# 2. Save the uniast for the first testdata item (0_*) in each language to out/
+# $ OUTDIR=out/ ./script/run_testdata.sh first
+#
+# 3. Use a custom abcoder executable
+# OUTDIR=out/ ABCEXE="./other_abcoder" ./script/run_testdata.sh all
+
+if [[ "$1" != "all" && "$1" != "first" ]]; then
+	echo "Usage: $0 all|first [--dry-run|-n]" >&2
+	echo "	all:     Run on all testdata." >&2
+	echo "	first: Run only on testdata starting with '0_*' in each language directory." >&2
+	echo "	--dry-run|-n: Print commands without executing them." >&2
+	exit 1
+fi
+
+MODE=$1
+DRY_RUN=false
+if [[ "$2" == "--dry-run" || "$2" == "-n" ]]; then
+       DRY_RUN=true
+fi
+
+SCRIPT_DIR=$(dirname "$(readlink -f "$0")")
+REPO_ROOT=$(realpath --relative-to=$(pwd) "$SCRIPT_DIR/..")
+
+ABCEXE=${ABCEXE:-"$REPO_ROOT/abcoder"}
+OUTDIR=${OUTDIR:?Error: OUTDIR is a mandatory environment variable}
+PARALLEL_FLAGS=${PARALLEL_FLAGS:---tag}
+LANGS=${LANGS:-"go rust python cxx"}
+
+detect_jobs() {
+	local ABCEXE=${1:-$ABCEXE}
+	for lang in ${LANGS[@]}; do
+	local repo_glob="$REPO_ROOT/testdata/$lang/*"
+	if [[ "$MODE" == "first" ]]; then
+		repo_glob="$REPO_ROOT/testdata/$lang/0_*"
+	fi
+	for repo in $repo_glob; do
+		# Skip if glob doesn't match anything to avoid errors
+		[[ -e "$repo" ]] || continue
+		local rel_path=$(realpath --no-symlinks --relative-to="$REPO_ROOT/testdata" "$repo")
+		local outname=$(echo "$rel_path" | sed 's/[/:? ]/_/g')
+		echo $ABCEXE parse $lang $repo -o $OUTDIR/$outname.json
+		done
+	done
+}
+
+if [[ ! -x "$ABCEXE" ]]; then
+	echo "Error: The specified abcoder executable '$ABCEXE' does not exist or is not executable." >&2
+	exit 1
+fi
+mkdir -pv "$OUTDIR"
+detect_jobs
+if $DRY_RUN ; then
+	exit 0
+fi
+echo
+detect_jobs | parallel $PARALLEL_FLAGS -j$(nproc --all) --jobs 0 "eval {}" 2>&1
+
+echo
+echo "Verifying that all expected output files were generated..."
+all_files_exist=true
+# Rerun detect_jobs to get the list of expected json files and check for their existence.
+for file_path in $(detect_jobs | awk '{print $NF}'); do
+    if [[ ! -f "$file_path" ]]; then
+        echo "Error: Expected output file does not exist: $file_path" >&2
+        all_files_exist=false
+    fi
+done
+
+if [[ "$all_files_exist" == "false" ]]; then
+    echo "One or more output files are missing. Failing." >&2
+    exit 1
+else
+    echo "All expected output files were successfully generated."
+fi

script/run_testdata.sh

@@ -0,0 +1 @@
+../../ts-parser/test-repo