Redesign: 2 tabs (leaderboard + submit) with HF login

README.md

@@ -8,13 +8,15 @@ sdk_version: 5.50.0
 python_version: "3.10"
 app_file: app.py
 pinned: false
+hf_oauth: true
 ---
 
 # PRIMO — Patient Representations in Multi-Omics
 
 **A blind benchmark for transcriptomic foundation models.**
 
-Embed **every** dataset with your model and upload **one** file. A fixed linear
+The Space has two tabs: a **Leaderboard** and a **Submit** form (sign in with
+Hugging Face). Embed **every** dataset with your model and upload **one** file. A fixed linear
 probe grades each hidden task (a dataset may be scored on several targets) over
 its frozen folds, and the scores roll up into a per-specialty skill leaderboard.
 The datasets are opaque (`d001`, `d002`…) — you never see the disease/tissue or
@@ -64,6 +66,8 @@ python evaluator.py --submission my_embeddings.parquet
 
 ## Space configuration
 
+- **`hf_oauth: true`** (set above) turns on the Submit tab's *Sign in with
+  Hugging Face* button; submitting requires a logged-in HF account.
 - Set an **`HF_TOKEN`** Space secret (fine-grained) with: **read** on
   `ScientaLab/primo` (the public `datasets.yaml` manifest) and
   `ScientaLab/primo-labels` (the private `tasks.yaml` registry +
@@ -72,3 +76,6 @@ python evaluator.py --submission my_embeddings.parquet
 - Results persist as one normalized `task_results.csv` (`model_name, task_id,
   score, submitted_at`) in the results dataset; the leaderboard is recomputed
   from it by joining the registry, so it survives Space restarts.
+- Submitter contact metadata (HF username, email, paper / model links, notes)
+  persists to a separate `submissions.csv` in the same **private** results
+  dataset — it never reaches the public leaderboard.

app.py

@@ -34,30 +34,35 @@ from scoring import TaskScore, aggregate, public_facets
 
 TOKEN = os.environ.get("HF_TOKEN")
 RESULTS_FILE = "task_results.csv"
+SUBMISSIONS_FILE = "submissions.csv"
 RESULT_COLUMNS = ["model_name", "task_id", "score", "submitted_at"]
+SUBMISSION_COLUMNS = [
+    "model_name",
+    "submitted_at",
+    "hf_username",
+    "email",
+    "paper_link",
+    "hf_model_link",
+    "notes",
+]
 BASE_COLUMNS = ["model_name", "overall_skill", "submitted_at"]
 DETAIL_COLUMNS = ["dataset_id", "status", "skill"]
 
 
-def _read_results() -> pd.DataFrame:
+def _read_csv(filename: str, columns: list[str]) -> pd.DataFrame:
     from huggingface_hub import hf_hub_download
     from huggingface_hub.utils import EntryNotFoundError, RepositoryNotFoundError
 
     try:
-        path = hf_hub_download(
-            RESULTS_REPO, RESULTS_FILE, repo_type="dataset", token=TOKEN
-        )
+        path = hf_hub_download(RESULTS_REPO, filename, repo_type="dataset", token=TOKEN)
     except (RepositoryNotFoundError, EntryNotFoundError):
-        return pd.DataFrame(columns=RESULT_COLUMNS)
+        return pd.DataFrame(columns=columns)
     return pd.read_csv(path)
 
 
-def _append_results(rows: list[dict]) -> None:
+def _upload_csv(filename: str, df: pd.DataFrame) -> None:
     from huggingface_hub import HfApi
 
-    if not rows:
-        return
-    df = pd.concat([_read_results(), pd.DataFrame(rows)], ignore_index=True)
     api = HfApi(token=TOKEN)
     api.create_repo(RESULTS_REPO, repo_type="dataset", private=True, exist_ok=True)
     buffer = io.BytesIO()
@@ -65,12 +70,32 @@ def _append_results(rows: list[dict]) -> None:
     buffer.seek(0)
     api.upload_file(
         path_or_fileobj=buffer,
-        path_in_repo=RESULTS_FILE,
+        path_in_repo=filename,
         repo_id=RESULTS_REPO,
         repo_type="dataset",
     )
 
 
+def _read_results() -> pd.DataFrame:
+    return _read_csv(RESULTS_FILE, RESULT_COLUMNS)
+
+
+def _append_results(rows: list[dict]) -> None:
+    if not rows:
+        return
+    df = pd.concat([_read_results(), pd.DataFrame(rows)], ignore_index=True)
+    _upload_csv(RESULTS_FILE, df)
+
+
+def _append_submission(meta: dict) -> None:
+    """Persist a submitter's contact metadata to the private results repo."""
+    df = pd.concat(
+        [_read_csv(SUBMISSIONS_FILE, SUBMISSION_COLUMNS), pd.DataFrame([meta])],
+        ignore_index=True,
+    )
+    _upload_csv(SUBMISSIONS_FILE, df)
+
+
 def _round(value: float | None) -> float | None:
     return round(value, 4) if value is not None else None
 
@@ -211,12 +236,24 @@ def _detail_table(result: dict) -> pd.DataFrame:
     return table[DETAIL_COLUMNS]
 
 
-def evaluate(submission_path: str, model_name: str):
+def evaluate(
+    submission_path: str,
+    model_name: str,
+    email: str,
+    paper_link: str,
+    hf_model_link: str,
+    notes: str,
+    profile: gr.OAuthProfile | None,
+):
     empty = pd.DataFrame(columns=DETAIL_COLUMNS)
+    if profile is None:
+        return "Please sign in with Hugging Face to submit.", leaderboard(), empty
     if not submission_path:
         return "Please upload a submission file.", leaderboard(), empty
     if not model_name or not model_name.strip():
         return "Please enter a model name.", leaderboard(), empty
+    if not email or not email.strip():
+        return "Please enter a contact email.", leaderboard(), empty
     try:
         result = score_all(submission_path, TOKEN)
     except SubmissionError as error:
@@ -251,8 +288,18 @@ def evaluate(submission_path: str, model_name: str):
             }
             for task in result["per_task"]
         ]
+        meta = {
+            "model_name": model,
+            "submitted_at": submitted_at,
+            "hf_username": profile.username,
+            "email": email.strip(),
+            "paper_link": (paper_link or "").strip(),
+            "hf_model_link": (hf_model_link or "").strip(),
+            "notes": (notes or "").strip(),
+        }
         try:
             _append_results(rows)
+            _append_submission(meta)
         except Exception as error:  # noqa: BLE001
             summary += f"\n\n⚠️ scored, but leaderboard not saved: {error}"
     return summary, leaderboard(), detail
@@ -261,32 +308,74 @@ def evaluate(submission_path: str, model_name: str):
 def build_demo() -> gr.Blocks:
     with gr.Blocks(title="PRIMO Benchmark") as demo:
         gr.Markdown(
-            "# 🧬 PRIMO\n\n"
+            "# 🧬 PRIMO — Patient Representations in Multi-Omics\n\n"
             "A **blind benchmark for transcriptomic foundation models** — grade "
             "your model's patient-level embeddings against real clinical signal, "
-            "without ever seeing the labels.\n\n"
-            "**Submit in 3 steps:**\n"
-            "1. **Get the data** → download the opaque datasets from "
-            "[ScientaLab/primo](https://huggingface.co/datasets/ScientaLab/primo)"
-            " (start with its `datasets.yaml`).\n"
-            "2. **Embed every dataset** → build **one** file: `dataset_id`, "
-            "`sample_id`, then one column per embedding dim (`e0`, `e1`, …). "
-            "CSV / TSV / Parquet, or NPZ.\n"
-            "3. **Upload below**, name your model, and hit **Evaluate**.\n\n"
-            "A fixed linear probe scores each hidden task (AUROC or Pearson), "
-            "rescaled to a 0–1 skill and rolled up per specialty. **Only "
-            "submissions covering every dataset are ranked.**"
+            "without ever seeing the labels."
         )
-        model_tb = gr.Textbox(label="Model name", placeholder="e.g. eva-rna-v1")
-        file_in = gr.File(
-            label="Submission (.csv / .tsv / .parquet / .npz)", type="filepath"
+        with gr.Tabs():
+            with gr.Tab("🏆 Leaderboard"):
+                gr.Markdown(
+                    "Overall skill plus a skill per medical specialty, on a 0–1 "
+                    "scale (0 = chance, 1 = perfect). **Only submissions covering "
+                    "every dataset are ranked.**"
+                )
+                board = gr.Dataframe(label="Leaderboard", interactive=False)
+            with gr.Tab("📤 Submit"):
+                gr.Markdown(
+                    "# Model submission\n\n"
+                    "1. **Get the data** → download the opaque datasets from "
+                    "[ScientaLab/primo](https://huggingface.co/"
+                    "datasets/ScientaLab/primo) (start with its `datasets.yaml`).\n"
+                    "2. **Embed every dataset** → build **one** file: `dataset_id`, "
+                    "`sample_id`, then one column per embedding dim (`e0`, `e1`, …). "
+                    "CSV / TSV / Parquet, or NPZ.\n"
+                    "3. **Sign in, fill the form, and hit Evaluate.** A fixed linear "
+                    "probe scores each hidden task (AUROC or Pearson), rescaled to "
+                    "a 0–1 skill and rolled up per specialty."
+                )
+                gr.LoginButton()
+                with gr.Row():
+                    with gr.Column():
+                        model_tb = gr.Textbox(
+                            label="Model name",
+                            placeholder="e.g. eva-rna-v1",
+                            info="Shown on the leaderboard.",
+                        )
+                        email_tb = gr.Textbox(
+                            label="Email address",
+                            placeholder="you@lab.org",
+                            info="Contact for this submission — kept private.",
+                        )
+                        notes_tb = gr.Textbox(
+                            label="Training data / notes (optional)",
+                            placeholder="e.g. pretrained on atlas X",
+                            info="About the model or its training data.",
+                        )
+                    with gr.Column():
+                        paper_tb = gr.Textbox(
+                            label="Paper link (optional)",
+                            placeholder="https://arxiv.org/abs/...",
+                        )
+                        hf_tb = gr.Textbox(
+                            label="Hugging Face model link (optional)",
+                            placeholder="https://huggingface.co/...",
+                        )
+                        file_in = gr.File(
+                            label="Submission (.csv / .tsv / .parquet / .npz)",
+                            type="filepath",
+                        )
+                run_btn = gr.Button("Evaluate", variant="primary")
+                result_md = gr.Markdown()
+                detail = gr.Dataframe(
+                    label="Your submission — per dataset", interactive=False
+                )
+
+        run_btn.click(
+            evaluate,
+            [file_in, model_tb, email_tb, paper_tb, hf_tb, notes_tb],
+            [result_md, board, detail],
         )
-        run_btn = gr.Button("Evaluate", variant="primary")
-        result_md = gr.Markdown()
-        board = gr.Dataframe(label="Leaderboard", interactive=False)
-        detail = gr.Dataframe(label="Your submission — per dataset", interactive=False)
-
-        run_btn.click(evaluate, [file_in, model_tb], [result_md, board, detail])
         demo.load(leaderboard, None, board)
     return demo
 

requirements.txt

@@ -3,5 +3,5 @@ pandas
 scikit-learn
 pyyaml
 huggingface_hub<1.0
-gradio==5.50.0
+gradio[oauth]==5.50.0
 audioop-lts; python_version >= "3.13"