Dress/.github/workflows/check_exif.yml

name: Check Image EXIF Data

on:
  pull_request_target:
    types: [opened, synchronize, reopened]
    paths: ['**/*.jpg', '**/*.jpeg', '**/*.png', '**/*.webp', '**/*.tiff', '**/*.tif', '**/*.heic', '**/*.heif', '**/*.avif', '**/*.gif']

permissions:
  contents: read
  pull-requests: write
  issues: write

concurrency:
  group: exif-pr-${{ github.event.pull_request.number }}
  cancel-in-progress: true

jobs:
  check-exif:
    runs-on: ubuntu-latest
    steps:
      - name: Install exiftool
        run: |
          command -v exiftool && exit 0
          sudo apt-get install -y --no-install-recommends libimage-exiftool-perl \
            || (sudo apt-get update && sudo apt-get install -y --no-install-recommends libimage-exiftool-perl)

      - name: Check EXIF in changed images
        uses: actions/github-script@v8
        with:
          script: |
            const { execSync } = require('child_process');
            const fs = require('fs');
            const path = require('path');
            const os = require('os');
            const pr = context.payload.pull_request;
            const { owner, repo } = context.repo;
            const prNumber = pr.number;
            const IMAGE_EXT = /\.(jpe?g|png|webp|tiff?|hei[cf]|avif|gif)$/i;

            // 1. Collect changed image files
            // Use compareCommitsWithBasehead (base.sha...head.sha) instead of pulls.listFiles.
            // pulls.listFiles computes the diff against the merge base, which can be a very old ancestor
            // when a contributor's fork is far behind main (e.g. after clicking "Update branch" which
            // creates a merge commit). This causes hundreds of unrelated files to appear as "changed".
            // compareCommitsWithBasehead diffs exactly what this PR's head adds on top of main's current
            // HEAD, regardless of the fork's branch history.
            const changedImages = [];
            let comparePage = 1;
            while (true) {
              const { data: cmp } = await github.rest.repos.compareCommitsWithBasehead({
                owner, repo,
                basehead: `${pr.base.sha}...${pr.head.sha}`,
                per_page: 100, page: comparePage,
              });
              for (const f of (cmp.files || [])) {
                if (['added', 'modified', 'renamed', 'copied'].includes(f.status) && IMAGE_EXT.test(f.filename))
                  changedImages.push(f.filename);
              }
              if (!cmp.files || cmp.files.length < 100) break;
              comparePage++;
            }
            if (!changedImages.length) { core.info('No changed images.'); return; }
            core.info(`Checking ${changedImages.length} image(s) for EXIF data.`);

            // 2. Download & check each image
            const tmp = fs.mkdtempSync(path.join(os.tmpdir(), 'exif-'));
            const exifFiles = [];

            for (const file of changedImages) {
              const tmpFile = path.join(tmp, file.replace(/[\\/]/g, '__'));
              try {
                // '#' must be percent-encoded in the API path; otherwise GitHub treats it as a URL fragment.
                const apiPath = file.replace(/#/g, '%23');
                const { data } = await github.rest.repos.getContent({
                  owner: pr.head.repo.owner.login, repo: pr.head.repo.name,
                  path: apiPath, ref: pr.head.sha
                });
                let raw;
                if (!Array.isArray(data) && data.content) raw = Buffer.from(data.content, 'base64');
                else if (!Array.isArray(data) && data.download_url) {
                  const resp = await fetch(data.download_url);
                  if (!resp.ok) throw new Error(`Download ${resp.status}`);
                  raw = Buffer.from(await resp.arrayBuffer());
                } else throw new Error('Cannot fetch file content');

                fs.writeFileSync(tmpFile, raw);
                // Check only high-sensitivity fields:
                //   GPS:*               — precise location coordinates
                //   IPTC address fields — textual shooting address
                //   XMP location/contact fields — address & contact info written by editing software
                const HIGH_SENS_ARGS = [
                  '-GPS:GPSLatitude', '-GPS:GPSLongitude', '-GPS:GPSAltitude',
                  '-GPS:GPSDateStamp', '-GPS:GPSTimeStamp',
                  '-GPS:GPSSpeed', '-GPS:GPSTrack', '-GPS:GPSImgDirection',
                  '-IPTC:City', '-IPTC:Country-PrimaryLocationName',
                  '-IPTC:Sub-location', '-IPTC:Province-State',
                  '-XMP-photoshop:City', '-XMP-photoshop:Country', '-XMP-photoshop:State',
                  '-XMP-iptcCore:CreatorWorkEmail', '-XMP-iptcCore:CreatorWorkTelephone',
                  '-XMP-iptcCore:CreatorCity', '-XMP-iptcCore:CreatorCountry',
                  '-XMP-iptcCore:CreatorPostalCode',
                ];
                const safeFile = tmpFile.replace(/'/g, "'\\\\'");
                // Filter out empty/undef/all-zero placeholder values (e.g. GPS IFD container with no real coords).
                // `: *(undef|00:00:00)?$` matches lines like "GPSLatitude: ", "GPSAltitude: undef", "GPSTimeStamp: 00:00:00"
                const count = parseInt(execSync(
                  `exiftool ${HIGH_SENS_ARGS.join(' ')} -S -q -- '${safeFile}' 2>/dev/null | grep -Ev ': *(undef|00:00:00)?$' | wc -l`,
                  { encoding: 'utf-8', timeout: 10000 }
                ).trim()) || 0;
                if (count > 0) { exifFiles.push(file); core.info(`Sensitive EXIF found: ${file} (${count} tags)`); }
              } catch (e) {
                // Download failures (e.g. diverged fork, encoding issues) are non-fatal.
                // Skip the file with a warning rather than failing the entire check.
                core.warning(`Skipping ${file}: ${e.message}`);
              }
              finally { try { fs.unlinkSync(tmpFile); } catch {} }
            }
            try { fs.rmdirSync(tmp); } catch {}

            // 3. Report results
            if (exifFiles.length) {
              const fileList = exifFiles.map(f => `- ${f}`).join('\n');
              const body = [
                '我们检测到以下文件包含**高敏感 EXIF 信息**（如 GPS 坐标、地址或联系方式），请移除后再次提交:',
                'We detected **high-sensitivity EXIF data** (e.g. GPS coordinates, address, or contact fields) in the following files. Please remove them and resubmit:',
                '',
                fileList,
                '',
                '**仅高敏感字段（GPS / 地址 / 联系方式）需要移除，普通摄影参数（光圈、快门等）无需处理。**',
                '**Only high-sensitivity fields (GPS / address / contact info) need to be removed. Regular photography parameters (aperture, shutter speed, etc.) are fine to keep.**',
                '',
                `📖 [EXIF 说明 / EXIF Guide](https://github.com/${owner}/${repo}/blob/master/EXIF.md) · [CONTRIBUTING.md](https://github.com/${owner}/${repo}/blob/master/CONTRIBUTING.md)`
              ].join('\n');
              try { await github.rest.issues.createComment({ owner, repo, issue_number: prNumber, body }); }
              catch (e) { core.warning(`Cannot comment: ${e.message}. Check repo Settings > Actions > Workflow permissions.`); }
              core.setFailed('EXIF data detected. Please remove EXIF data for privacy protection.');
            } else {
              core.info('No EXIF data found. ✅');
            }