Auth

Dockerfile

@@ -46,6 +46,7 @@ COPY --chown=uvuser:uvuser . ./
 # Expose Gradio app port
 EXPOSE 7860
 ENV GRADIO_SERVER_NAME="0.0.0.0"
+ENV SYSTEM=spaces
 
 # Run the application
-CMD ["uv", "run", "python", "app.py"]
+CMD ["uv", "run", "python", "yourbench_space/app.py"]

pyproject.toml

@@ -19,7 +19,7 @@ dependencies = [
     "torch>=2.6.0",
     "tqdm>=4.67.1",
     "transformers>=4.48.3",
-    "gradiologin>=0.1.0",
+    "ruamel-yaml>=0.18.10",
 ]
 
 [build-system]

uv.lock

@@ -532,20 +532,6 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/95/cb/002424d4f5af1425f9cfe7dcee3ed795ed1367bf0f185a6c4bf81385e1d6/gradio_client-1.7.2-py3-none-any.whl", hash = "sha256:50d61b4db3e87639430a121a7cde4303055486ed72a5035edae94b4fbe6a0e6b", size = 322052 },
 ]
 
-[[package]]
-name = "gradiologin"
-version = "0.1.0"
-source = { registry = "https://pypi.org/simple" }
-dependencies = [
-    { name = "authlib" },
-    { name = "gradio" },
-    { name = "itsdangerous" },
-]
-sdist = { url = "https://files.pythonhosted.org/packages/75/88/ca2122597e416e87bbc97201909fcc028f3296b39d6b5b20409d7583d204/gradiologin-0.1.0.tar.gz", hash = "sha256:388fcfb94161fc1a61492fc729bc4dff3e437b8ac063735f5a9e8886b36f35b1", size = 5905 }
-wheels = [
-    { url = "https://files.pythonhosted.org/packages/fc/55/c4c397bb0f9a512e9a77daaf49dbdebeaa2487930b23caa62a716ef4bf6e/gradiologin-0.1.0-py3-none-any.whl", hash = "sha256:111af87a94867f9ca0602976fdbe2f23be8d5557c91b3a4e5c2db096e7f053d0", size = 7238 },
-]
-
 [[package]]
 name = "groovy"
 version = "0.1.2"
@@ -1503,6 +1489,35 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/a3/4f/0fce63e0f5cdd658e71e21abd17ac1bc9312741ebb8b3f74eeed2ebdf771/rpds_py-0.23.1-cp312-cp312-win_amd64.whl", hash = "sha256:b03a8d50b137ee758e4c73638b10747b7c39988eb8e6cd11abb7084266455165", size = 237426 },
 ]
 
+[[package]]
+name = "ruamel-yaml"
+version = "0.18.10"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "ruamel-yaml-clib", marker = "platform_python_implementation == 'CPython'" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/ea/46/f44d8be06b85bc7c4d8c95d658be2b68f27711f279bf9dd0612a5e4794f5/ruamel.yaml-0.18.10.tar.gz", hash = "sha256:20c86ab29ac2153f80a428e1254a8adf686d3383df04490514ca3b79a362db58", size = 143447 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/c2/36/dfc1ebc0081e6d39924a2cc53654497f967a084a436bb64402dfce4254d9/ruamel.yaml-0.18.10-py3-none-any.whl", hash = "sha256:30f22513ab2301b3d2b577adc121c6471f28734d3d9728581245f1e76468b4f1", size = 117729 },
+]
+
+[[package]]
+name = "ruamel-yaml-clib"
+version = "0.2.12"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/20/84/80203abff8ea4993a87d823a5f632e4d92831ef75d404c9fc78d0176d2b5/ruamel.yaml.clib-0.2.12.tar.gz", hash = "sha256:6c8fbb13ec503f99a91901ab46e0b07ae7941cd527393187039aec586fdfd36f", size = 225315 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/48/41/e7a405afbdc26af961678474a55373e1b323605a4f5e2ddd4a80ea80f628/ruamel.yaml.clib-0.2.12-cp312-cp312-macosx_14_0_arm64.whl", hash = "sha256:20b0f8dc160ba83b6dcc0e256846e1a02d044e13f7ea74a3d1d56ede4e48c632", size = 133433 },
+    { url = "https://files.pythonhosted.org/packages/ec/b0/b850385604334c2ce90e3ee1013bd911aedf058a934905863a6ea95e9eb4/ruamel.yaml.clib-0.2.12-cp312-cp312-manylinux2014_aarch64.whl", hash = "sha256:943f32bc9dedb3abff9879edc134901df92cfce2c3d5c9348f172f62eb2d771d", size = 647362 },
+    { url = "https://files.pythonhosted.org/packages/44/d0/3f68a86e006448fb6c005aee66565b9eb89014a70c491d70c08de597f8e4/ruamel.yaml.clib-0.2.12-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:95c3829bb364fdb8e0332c9931ecf57d9be3519241323c5274bd82f709cebc0c", size = 754118 },
+    { url = "https://files.pythonhosted.org/packages/52/a9/d39f3c5ada0a3bb2870d7db41901125dbe2434fa4f12ca8c5b83a42d7c53/ruamel.yaml.clib-0.2.12-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:749c16fcc4a2b09f28843cda5a193e0283e47454b63ec4b81eaa2242f50e4ccd", size = 706497 },
+    { url = "https://files.pythonhosted.org/packages/b0/fa/097e38135dadd9ac25aecf2a54be17ddf6e4c23e43d538492a90ab3d71c6/ruamel.yaml.clib-0.2.12-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:bf165fef1f223beae7333275156ab2022cffe255dcc51c27f066b4370da81e31", size = 698042 },
+    { url = "https://files.pythonhosted.org/packages/ec/d5/a659ca6f503b9379b930f13bc6b130c9f176469b73b9834296822a83a132/ruamel.yaml.clib-0.2.12-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:32621c177bbf782ca5a18ba4d7af0f1082a3f6e517ac2a18b3974d4edf349680", size = 745831 },
+    { url = "https://files.pythonhosted.org/packages/db/5d/36619b61ffa2429eeaefaab4f3374666adf36ad8ac6330d855848d7d36fd/ruamel.yaml.clib-0.2.12-cp312-cp312-musllinux_1_2_aarch64.whl", hash = "sha256:b82a7c94a498853aa0b272fd5bc67f29008da798d4f93a2f9f289feb8426a58d", size = 715692 },
+    { url = "https://files.pythonhosted.org/packages/b1/82/85cb92f15a4231c89b95dfe08b09eb6adca929ef7df7e17ab59902b6f589/ruamel.yaml.clib-0.2.12-cp312-cp312-win32.whl", hash = "sha256:e8c4ebfcfd57177b572e2040777b8abc537cdef58a2120e830124946aa9b42c5", size = 98777 },
+    { url = "https://files.pythonhosted.org/packages/d7/8f/c3654f6f1ddb75daf3922c3d8fc6005b1ab56671ad56ffb874d908bfa668/ruamel.yaml.clib-0.2.12-cp312-cp312-win_amd64.whl", hash = "sha256:0467c5965282c62203273b838ae77c0d29d7638c8a4e3a1c8bdd3602c10904e4", size = 115523 },
+]
+
 [[package]]
 name = "ruff"
 version = "0.9.9"
@@ -1971,7 +1986,6 @@ dependencies = [
     { name = "asyncio" },
     { name = "datasets" },
     { name = "gradio", extra = ["oauth"] },
-    { name = "gradiologin" },
     { name = "hf-transfer" },
     { name = "langfuse" },
     { name = "litellm" },
@@ -1980,6 +1994,7 @@ dependencies = [
     { name = "matplotlib" },
     { name = "openai" },
     { name = "python-dotenv" },
+    { name = "ruamel-yaml" },
     { name = "torch" },
     { name = "tqdm" },
     { name = "transformers" },
@@ -1991,7 +2006,6 @@ requires-dist = [
     { name = "asyncio", specifier = ">=3.4.3" },
     { name = "datasets", specifier = ">=3.3.0" },
     { name = "gradio", extras = ["oauth"], specifier = ">=5.20.0" },
-    { name = "gradiologin", specifier = ">=0.1.0" },
     { name = "hf-transfer", specifier = ">=0.1.9" },
     { name = "langfuse", specifier = ">=2.59.3" },
     { name = "litellm", specifier = ">=1.61.16" },
@@ -2000,6 +2014,7 @@ requires-dist = [
     { name = "matplotlib", specifier = ">=3.10.0" },
     { name = "openai", specifier = ">=1.63.0" },
     { name = "python-dotenv", specifier = ">=1.0.1" },
+    { name = "ruamel-yaml", specifier = ">=0.18.10" },
     { name = "torch", specifier = ">=2.6.0" },
     { name = "tqdm", specifier = ">=4.67.1" },
     { name = "transformers", specifier = ">=4.48.3" },

app.py → yourbench_space/app.py

@@ -8,14 +8,15 @@ import shutil
 import io
 import yaml
 import gradio as gr
-import gradiologin as gl
 from huggingface_hub import whoami
 from loguru import logger
 from yourbench.pipeline import run_pipeline
 
-UPLOAD_DIRECTORY = pathlib.Path("/app/uploaded_files")
+from yourbench_space.config import generate_config, save_config
+from yourbench_space.utils import CONFIG_PATH, UPLOAD_DIRECTORY
+
 UPLOAD_DIRECTORY.mkdir(parents=True, exist_ok=True)
-CONFIG_PATH = pathlib.Path("/app/yourbench_config.yml")
+
 
 logger.remove()
 logger.add(sys.stderr, level="INFO")
@@ -88,108 +89,6 @@ command = ["uv", "run", "yourbench", f"--config={CONFIG_PATH}"]
 manager = SubprocessManager(command)
 
 
-def generate_config(hf_token: gr.OAuthToken | None, hf_org, model_name, provider, base_url, api_key, max_concurrent_requests):
-    config = {
-        "hf_configuration": {
-            "token": hf_token,
-            "private": True,
-            "hf_organization": hf_org
-        },
-        "model_list": [{
-            "model_name": model_name,
-            "provider": provider,
-            "base_url": base_url,
-            "api_key": api_key,
-            "max_concurrent_requests": max_concurrent_requests
-        }],
-        "model_roles": {role: [model_name] for role in [
-            "ingestion", "summarization", "single_shot_question_generation",
-            "multi_hop_question_generation", "answer_generation", "judge_answers"
-        ]},
-        "inference_config": {"max_concurrent_requests": 16},
-        "pipeline": {
-            "ingestion": {
-                "source_documents_dir": "/app/uploaded_files",
-                "output_dir": "/app/ingested",
-                "run": True
-            },
-            "upload_ingest_to_hub": {
-                "source_documents_dir": "/app/ingested",
-                "hub_dataset_name": "test_ingested_documents",
-                "local_dataset_path": "/app/ingested_dataset",
-                "run": True
-            },
-            "summarization": {
-                "source_dataset_name": "test_ingested_documents",
-                "output_dataset_name": "test_summaries",
-                "local_dataset_path": "/results/test_summaries",
-                "concat_existing_dataset": False,
-                "run": True
-            },
-            "chunking": {
-                "source_dataset_name": "test_summaries",
-                "output_dataset_name": "test_chunked_documents",
-                "local_dataset_path": "/results/test_chunked_documents",
-                "concat_existing_dataset": False,
-                "chunking_configuration": {
-                    "l_min_tokens": 64,
-                    "l_max_tokens": 128,
-                    "tau_threshold": 0.3,
-                    "h_min": 2,
-                    "h_max": 4
-                },
-                "run": True
-            },
-            "single_shot_question_generation": {
-                "source_dataset_name": "test_chunked_documents",
-                "output_dataset_name": "test_single_shot_questions",
-                "local_dataset_path": "/results/test_single_shot_questions",
-                "diversification_seed": "24 year old adult",
-                "concat_existing_dataset": False,
-                "run": True
-            },
-            "multi_hop_question_generation": {
-                "source_dataset_name": "test_chunked_documents",
-                "output_dataset_name": "test_multi_hop_questions",
-                "local_dataset_path": "/results/test_multi_hop_questions",
-                "concat_existing_dataset": False,
-                "run": True
-            },
-            "answer_generation": {
-                "run": True,
-                "question_dataset_name": "test_single_shot_questions",
-                "output_dataset_name": "test_answered_questions",
-                "local_dataset_path": "/results/test_answered_questions",
-                "concat_existing_dataset": False,
-                "strategies": [{
-                    "name": "zeroshot",
-                    "prompt": "ZEROSHOT_QA_USER_PROMPT",
-                    "model_name": model_name
-                }, {
-                    "name": "gold",
-                    "prompt": "GOLD_QA_USER_PROMPT",
-                    "model_name": model_name
-                }]
-            },
-            "judge_answers": {
-                "run": True,
-                "source_judge_dataset_name": "test_answered_questions",
-                "output_judged_dataset_name": "test_judged_comparisons",
-                "local_dataset_path": "/results/test_judged_comparisons",
-                "concat_existing_dataset": False,
-                "comparing_strategies": [["zeroshot", "gold"]],
-                "chunk_column_index": 0,
-                "random_seed": 42
-            }
-        }
-    }
-    return yaml.dump(config, default_flow_style=False)
-
-def save_config(yaml_text):
-    with open(CONFIG_PATH, "w") as file:
-        file.write(yaml_text)
-    return "✅ Config saved!"
-
 def save_files(files: list[str]):
     saved_paths = [shutil.move(str(pathlib.Path(file)), str(UPLOAD_DIRECTORY / pathlib.Path(file).name)) for file in files]
     return f"Files saved to: {', '.join(saved_paths)}"
@@ -288,4 +187,4 @@ with gr.Blocks() as app:
         kill_button = gr.Button("Kill Task")
         kill_button.click(manager.kill_process)
 
-app.launch()
+app.launch()

yourbench_space/config.py

@@ -0,0 +1,107 @@
+import yaml
+import gradio as gr
+
+from yourbench_space.utils import CONFIG_PATH
+
+
+def generate_config(hf_token: gr.OAuthToken | None, hf_org, model_name, provider, base_url, api_key, max_concurrent_requests):
+    config = {
+        "hf_configuration": {
+            "token": hf_token,
+            "private": True,
+            "hf_organization": hf_org
+        },
+        "model_list": [{
+            "model_name": model_name,
+            "provider": provider,
+            "base_url": base_url,
+            "api_key": api_key,
+            "max_concurrent_requests": max_concurrent_requests
+        }],
+        "model_roles": {role: [model_name] for role in [
+            "ingestion", "summarization", "single_shot_question_generation",
+            "multi_hop_question_generation", "answer_generation", "judge_answers"
+        ]},
+        "inference_config": {"max_concurrent_requests": 16},
+        "pipeline": {
+            "ingestion": {
+                "source_documents_dir": "/app/uploaded_files",
+                "output_dir": "/app/ingested",
+                "run": True
+            },
+            "upload_ingest_to_hub": {
+                "source_documents_dir": "/app/ingested",
+                "hub_dataset_name": "test_ingested_documents",
+                "local_dataset_path": "/app/ingested_dataset",
+                "run": True
+            },
+            "summarization": {
+                "source_dataset_name": "test_ingested_documents",
+                "output_dataset_name": "test_summaries",
+                "local_dataset_path": "/results/test_summaries",
+                "concat_existing_dataset": False,
+                "run": True
+            },
+            "chunking": {
+                "source_dataset_name": "test_summaries",
+                "output_dataset_name": "test_chunked_documents",
+                "local_dataset_path": "/results/test_chunked_documents",
+                "concat_existing_dataset": False,
+                "chunking_configuration": {
+                    "l_min_tokens": 64,
+                    "l_max_tokens": 128,
+                    "tau_threshold": 0.3,
+                    "h_min": 2,
+                    "h_max": 4
+                },
+                "run": True
+            },
+            "single_shot_question_generation": {
+                "source_dataset_name": "test_chunked_documents",
+                "output_dataset_name": "test_single_shot_questions",
+                "local_dataset_path": "/results/test_single_shot_questions",
+                "diversification_seed": "24 year old adult",
+                "concat_existing_dataset": False,
+                "run": True
+            },
+            "multi_hop_question_generation": {
+                "source_dataset_name": "test_chunked_documents",
+                "output_dataset_name": "test_multi_hop_questions",
+                "local_dataset_path": "/results/test_multi_hop_questions",
+                "concat_existing_dataset": False,
+                "run": True
+            },
+            "answer_generation": {
+                "run": True,
+                "question_dataset_name": "test_single_shot_questions",
+                "output_dataset_name": "test_answered_questions",
+                "local_dataset_path": "/results/test_answered_questions",
+                "concat_existing_dataset": False,
+                "strategies": [{
+                    "name": "zeroshot",
+                    "prompt": "ZEROSHOT_QA_USER_PROMPT",
+                    "model_name": model_name
+                }, {
+                    "name": "gold",
+                    "prompt": "GOLD_QA_USER_PROMPT",
+                    "model_name": model_name
+                }]
+            },
+            "judge_answers": {
+                "run": True,
+                "source_judge_dataset_name": "test_answered_questions",
+                "output_judged_dataset_name": "test_judged_comparisons",
+                "local_dataset_path": "/results/test_judged_comparisons",
+                "concat_existing_dataset": False,
+                "comparing_strategies": [["zeroshot", "gold"]],
+                "chunk_column_index": 0,
+                "random_seed": 42
+            }
+        }
+    }
+    return yaml.dump(config, default_flow_style=False)
+
+def save_config(yaml_text):
+    with open(CONFIG_PATH, "w") as file:
+        file.write(yaml_text)
+    return "✅ Config saved!"

yourbench_space/utils.py

@@ -0,0 +1,4 @@
+import pathlib
+
+UPLOAD_DIRECTORY = pathlib.Path("/app/uploaded_files")
+CONFIG_PATH = pathlib.Path("/app/yourbench_config.yml")