Initial ArtiFinder integration

martonbognar · martonbognar · commit e13fea6d44ca · 2026-06-24T15:57:04.000+02:00
diff --git a/.github/workflows/deploy.yml b/.github/workflows/deploy.yml
@@ -0,0 +1,54 @@
+name: Deploy
+
+on:
+  push:
+    branches: [main]
+  workflow_dispatch:
+
+permissions:
+  contents: read
+  pages: write
+  id-token: write
+
+concurrency:
+  group: pages
+  cancel-in-progress: false
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          submodules: recursive
+
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.x"
+
+      - run: pip install pyyaml
+
+      - run: python3 generate_artifinder_data.py
+
+      - uses: ruby/setup-ruby@v1
+        with:
+          ruby-version: "3.2"
+          bundler-cache: true
+
+      - uses: actions/configure-pages@v5
+
+      - run: bundle exec jekyll build
+        env:
+          JEKYLL_ENV: production
+
+      - uses: actions/upload-pages-artifact@v3
+
+  deploy:
+    needs: build
+    runs-on: ubuntu-latest
+    environment:
+      name: github-pages
+      url: ${{ steps.deployment.outputs.page_url }}
+    steps:
+      - id: deployment
+        uses: actions/deploy-pages@v4
diff --git a/.gitignore b/.gitignore
@@ -1,5 +1,8 @@
 _site
 .jekyll-cache
+_data/artifinder_artifacts.yaml
+_data/artifinder_links.yaml
+_data/artifinder_index.yaml
 .jekyll-metadata
 .DS_Store
 vendor/
diff --git a/_conferences/ndss2024/results.md b/_conferences/ndss2024/results.md
@@ -180,6 +180,7 @@ artifacts:
       <th>Functional</th>
       <th>Reproduced</th>
       <th>Available at</th>
+      <th>Author's link</th>
     </tr>
   </thead>
   <tbody>
@@ -214,6 +215,7 @@ artifacts:
           <a href="{{artifact.artifact_url}}" target="_blank">Artifact</a><br>
         {% endif %}
       </td>
+      {% include artifinder_link_cell.html title=artifact.title %}
     </tr>
     {% endfor %}
   </tbody>
diff --git a/_conferences/ndss2025/results.md b/_conferences/ndss2025/results.md
@@ -282,6 +282,7 @@ artifacts:
       <th>Functional</th>
       <th>Reproduced</th>
       <th>Available at</th>
+      <th>Author's link</th>
     </tr>
   </thead>
   <tbody>
@@ -316,6 +317,7 @@ artifacts:
           <a href="{{artifact.artifact_url}}" target="_blank">Artifact</a><br>
         {% endif %}
       </td>
+      {% include artifinder_link_cell.html title=artifact.title %}
     </tr>
     {% endfor %}
   </tbody>
diff --git a/_conferences/usenixsec2020/results.md b/_conferences/usenixsec2020/results.md
@@ -195,6 +195,7 @@ artifacts:
       <th>Title</th>
       <th>Badge</th>
       <th>Artifact URL</th>
+      <th>Author's link</th>
     </tr>
   </thead>
   <tbody>
@@ -217,6 +218,7 @@ artifacts:
           <a href="{{artifact.artifact_url}}">Artifact</a>
         {% endif %}
       </td>
+      {% include artifinder_link_cell.html title=artifact.title %}
     </tr>
   {% endfor %}
   </tbody>
diff --git a/_conferences/usenixsec2021/results.md b/_conferences/usenixsec2021/results.md
@@ -176,6 +176,7 @@ artifacts:
       <th>Title</th>
       <th>Badge</th>
       <th>Artifact URL</th>
+      <th>Author's link</th>
     </tr>
   </thead>
   <tbody>
@@ -198,6 +199,7 @@ artifacts:
           <a href="{{artifact.artifact_url}}">Artifact</a>
         {% endif %}
       </td>
+      {% include artifinder_link_cell.html title=artifact.title %}
     </tr>
   {% endfor %}
   </tbody>
diff --git a/_conferences/usenixsec2022/results.md b/_conferences/usenixsec2022/results.md
@@ -693,6 +693,7 @@ Results include evaluation of the USENIX Security Summer submissions and will be
       <th>Repro.</th>
       <th>Repository</th>
       <th>Appendix</th>
+      <th>Author's link</th>
     </tr>
   </thead>
   <tbody>
@@ -728,6 +729,7 @@ Results include evaluation of the USENIX Security Summer submissions and will be
           <a href="{{artifact.appendix_url}}">Appendix</a>
         {% endif %}
       </td>
+      {% include artifinder_link_cell.html title=artifact.title %}
     </tr>
   {% endfor %}
   </tbody>
diff --git a/_conferences/usenixsec2023/results.md b/_conferences/usenixsec2023/results.md
@@ -881,6 +881,7 @@ artifacts:
       <th>Funct.</th>
       <th>Repro.</th>
       <th>Available At</th>
+      <th>Author's link</th>
     </tr>
   </thead>
   <tbody>
@@ -919,6 +920,7 @@ artifacts:
           <a href="{{artifact.appendix_url}}" target="_blank">Appendix</a><br>
         {% endif %}
       </td>
+      {% include artifinder_link_cell.html title=artifact.title %}
     </tr>
     {% endfor %}
   </tbody>
diff --git a/_conferences/usenixsec2024/results.md b/_conferences/usenixsec2024/results.md
@@ -1034,6 +1034,7 @@ artifacts:
       <th width="75px">Funct.</th>
       <th width="75px">Repro.</th>
       <th>Available At</th>
+      <th>Author's link</th>
     </tr>
   </thead>
   <tbody>
@@ -1079,6 +1080,7 @@ artifacts:
           📄 <a href="{{artifact.appendix_url}}" target="_blank">Appendix</a><br>
         {% endif %}
       </td>
+      {% include artifinder_link_cell.html title=artifact.title %}
     </tr>
   {% endfor %}
   </tbody>
diff --git a/_conferences/usenixsec2025/results.md b/_conferences/usenixsec2025/results.md
@@ -3127,6 +3127,7 @@ artifacts:
       <th width="75px">Funct.</th>
       <th width="75px">Repro.</th>
       <th>Available At</th>
+      <th>Author's link</th>
     </tr>
   </thead>
   <tbody>
@@ -3172,6 +3173,7 @@ artifacts:
           📄 <a href="{{artifact.appendix_url}}" target="_blank">Appendix</a><br>
         {% endif %}
       </td>
+      {% include artifinder_link_cell.html title=artifact.title %}
     </tr>
   {% endfor %}
   </tbody>
diff --git a/_includes/artifinder_link_cell.html b/_includes/artifinder_link_cell.html
@@ -0,0 +1,5 @@
+<td>
+  {% assign dir_name = page.path | remove_first:"_conferences/" | split:"/" | first %}
+  {% assign af_url = site.data.artifinder_links[dir_name][include.title] %}
+  {% if af_url %}<a href="{{ af_url }}" target="_blank">Author's link</a>{% endif %}
+</td>
diff --git a/_includes/masthead.html b/_includes/masthead.html
@@ -30,6 +30,9 @@
   {% endif %}
 {% endfor %}
 
+          <li class="masthead__menu-item">
+            <a href="{{ '/non-evaluated/' | relative_url }}">Non-evaluated</a>
+          </li>
         </ul>
         {% if site.search == true %}
         <button class="search__toggle" type="button">
diff --git a/generate_artifinder_data.py b/generate_artifinder_data.py
@@ -0,0 +1,168 @@
+#!/usr/bin/env python3
+"""Generate _data/artifinder_artifacts.yaml and _data/artifinder_links.yaml.
+
+artifinder_artifacts.yaml: non-evaluated papers with discovered artifacts,
+  cross-checked against _conferences to exclude formally evaluated ones.
+
+artifinder_links.yaml: for papers that went through evaluation, cases where
+  ArtiFinder found a different artifact URL than the one in results.md.
+"""
+
+import os
+import re
+import yaml
+
+DATA_DIR = "_artifinder/data"
+CONFERENCES_DIR = "_conferences"
+ARTIFACTS_OUT = "_data/artifinder_artifacts.yaml"
+LINKS_OUT = "_data/artifinder_links.yaml"
+
+# Maps _conferences directory prefix → _artifinder/data directory name
+CONF_PREFIX_MAP = {
+    "ndss": "ndss",
+    "usenixsec": "usenix",
+    "sp": "sp",
+}
+
+
+def normalize_title(title):
+    title = re.sub(r"\s*\[artifacts?\]", "", title, flags=re.IGNORECASE)
+    return re.sub(r"\s+", " ", title).strip().rstrip(".").lower()
+
+
+def normalize_url(url):
+    if not url:
+        return ""
+    url = url.strip().rstrip("/")
+    url = re.sub(r"^http://", "https://", url)
+    return url
+
+
+def read_front_matter(path):
+    with open(path) as f:
+        content = f.read().replace("\t", "  ")
+    parts = content.split("---")
+    if len(parts) < 3:
+        return {}
+    return yaml.safe_load(parts[1]) or {}
+
+
+def get_eval_url(artifact):
+    url = artifact.get("artifact_url")
+    if isinstance(url, list):
+        url = url[0] if url else None
+    return normalize_url(url or "")
+
+
+def load_conference_data():
+    """Return {dir_name: {norm_title: (raw_title, eval_url)}} for all matching conf-years."""
+    result = {}
+    for entry in os.listdir(CONFERENCES_DIR):
+        m = re.match(r"^([a-z]+)(\d{4})$", entry)
+        if not m or m.group(1) not in CONF_PREFIX_MAP:
+            continue
+        results_path = os.path.join(CONFERENCES_DIR, entry, "results.md")
+        if not os.path.exists(results_path):
+            continue
+        fm = read_front_matter(results_path)
+        by_norm = {}
+        for a in fm.get("artifacts", []):
+            by_norm[normalize_title(a["title"])] = (a["title"], get_eval_url(a))
+        if by_norm:
+            result[entry] = (CONF_PREFIX_MAP[m.group(1)], m.group(2), by_norm)
+    return result
+
+
+conf_data = load_conference_data()
+
+# Build evaluated-title lookup for artifact filtering: {(conf, year): set_of_norm_titles}
+evaluated_lookup = {
+    (conf, year): set(by_norm.keys())
+    for _, (conf, year, by_norm) in conf_data.items()
+}
+print(f"Loaded evaluated titles for {len(evaluated_lookup)} conference-years")
+
+# ── Generate artifinder_artifacts.yaml ────────────────────────────────────────
+
+conferences = sorted(os.listdir(DATA_DIR))
+artifacts_result = {}
+total_kept = 0
+total_excluded = 0
+
+for conf in conferences:
+    conf_dir = os.path.join(DATA_DIR, conf)
+    if not os.path.isdir(conf_dir):
+        continue
+
+    years = sorted(
+        (f[:-5] for f in os.listdir(conf_dir) if f.endswith(".yaml")),
+        reverse=True,
+    )
+
+    conf_data_out = {}
+    for year in years:
+        path = os.path.join(conf_dir, f"{year}.yaml")
+        with open(path) as fh:
+            papers = yaml.safe_load(fh) or []
+
+        evaluated = evaluated_lookup.get((conf, year), set())
+
+        artifacts = []
+        for p in papers:
+            if not p.get("discovered_artifact"):
+                continue
+            if normalize_title(p["title"]) in evaluated:
+                total_excluded += 1
+                continue
+            artifacts.append({
+                "title": p["title"],
+                "page_link": p.get("page_link"),
+                "artifact": p["discovered_artifact"],
+            })
+
+        total_kept += len(artifacts)
+        if artifacts:
+            conf_data_out[year] = artifacts
+
+    if conf_data_out:
+        artifacts_result[conf] = conf_data_out
+
+with open(ARTIFACTS_OUT, "w") as fh:
+    yaml.dump(artifacts_result, fh, allow_unicode=True, sort_keys=False)
+
+print(f"artifinder_artifacts.yaml — excluded: {total_excluded}, kept: {total_kept}")
+
+# ── Generate artifinder_links.yaml ────────────────────────────────────────────
+# For evaluated papers where ArtiFinder found a different artifact URL.
+# Keyed by _conferences directory name → raw results.md title → AF url.
+
+links_result = {}
+total_links = 0
+
+for dir_name, (conf, year, by_norm) in sorted(conf_data.items()):
+    af_path = os.path.join(DATA_DIR, conf, f"{year}.yaml")
+    if not os.path.exists(af_path):
+        continue
+
+    with open(af_path) as fh:
+        af_papers = yaml.safe_load(fh) or []
+
+    af_by_norm = {
+        normalize_title(p["title"]): normalize_url(p.get("discovered_artifact") or "")
+        for p in af_papers
+    }
+
+    dir_links = {}
+    for norm_title, (raw_title, eval_url) in by_norm.items():
+        af_url = af_by_norm.get(norm_title, "")
+        if af_url and af_url != eval_url:
+            dir_links[raw_title] = af_url
+            total_links += 1
+
+    if dir_links:
+        links_result[dir_name] = dir_links
+
+with open(LINKS_OUT, "w") as fh:
+    yaml.dump(links_result, fh, allow_unicode=True, sort_keys=False)
+
+print(f"artifinder_links.yaml — {total_links} alternative links across {len(links_result)} conf-years")
diff --git a/index.md b/index.md
@@ -35,3 +35,8 @@ To increase the discoverability of these artifacts, the [ArtiFinder](https://git
 The following pages list the automatically extracted artifacts.
 As these results were not manually verified, mistakes are possible, for which pull requests are very welcome.
 Currently, ArtiFinder was used to collect artifacts from A* conferences from the years 2000--2025.
+
+{% for conf in site.data.artifinder_index %}
+**{{ conf[0] | upcase }}**: {% for year in conf[1] %}{{ year }} {% endfor %}
+
+{% endfor %}
diff --git a/non-evaluated.md b/non-evaluated.md
