update

app.py

@@ -1,5 +1,5 @@
-# dream_app.py
-import spaces
+# dreamon_app.py
+## this app is built based on https://huggingface.co/spaces/multimodalart/Dream/blob/main/app.py
 import torch
 import numpy as np
 import gradio as gr
@@ -431,7 +431,7 @@ def get_example_input():
     problem = problems[task_id]
     prefix, suffix = problem['prompt'], problem['suffix']
     test_case = problem['test']
-    return prefix, suffix, test_case, task_id
+    return prefix, '', suffix, test_case, task_id, ''
 
 def check_result(task_id, completion):
     # 从数据集中读取问题
@@ -458,55 +458,57 @@ def create_chatbot_demo():
             "[[Blog](https://hkunlp.github.io/blog/2025/dreamon/)]"
         )
 
-        with gr.Column():
-            # Prefix input
-            prefix_input = gr.Textbox(
-                label="Prefix Text",
-                placeholder="Enter the beginning of your text...",
-                lines=2
-            )
-            
-            # Middle generation/visualization area
-            output_vis = gr.Textbox(
-                label="Generated Text (Middle)", 
-                lines=2
-            )
+        with gr.Row():
+            with gr.Column():
+                # Prefix input
+                prefix_input = gr.Textbox(
+                    label="Prefix Text",
+                    placeholder="Enter the beginning of your text...",
+                    lines=2
+                )
+                
+                # Middle generation/visualization area
+                output_vis = gr.Textbox(
+                    label="Generated Text (Middle)", 
+                    lines=2
+                )
+                
+                # Suffix input
+                suffix_input = gr.Textbox(
+                    label="Suffix Text",
+                    placeholder="Enter the end of your text...",
+                    lines=2
+                )
+                
+                # Hidden Task ID input
+                task_id_input = gr.Textbox(
+                    label="Task ID",
+                    placeholder="Task ID will be stored here...",
+                    visible=False
+                )
             
-            # Suffix input
-            suffix_input = gr.Textbox(
-                label="Suffix Text",
-                placeholder="Enter the end of your text...",
-                lines=2
-            )
-
-            # Test Case input
-            test_case_input = gr.Textbox(
-                label="Test Case",
-                placeholder="Enter your test case here...",
-                lines=2
-            )
-
-            # Hidden Task ID input
-            task_id_input = gr.Textbox(
-                label="Task ID",
-                placeholder="Task ID will be stored here...",
-                visible=False
-            )
-
-            # Result of execution
-            result_output = gr.Textbox(
-                label="Result of Execution",
-                placeholder="Execution result will be shown here...",
-                lines=2
-            )
-
-            with gr.Row():
-                sample_btn = gr.Button("Example Prompt")
+            with gr.Column():
+                # Test Case input
+                test_case_input = gr.Textbox(
+                    label="Test Case",
+                    placeholder="Enter your test case here...",
+                    lines=2
+                )
+                
+                # Result of execution
+                result_output = gr.Textbox(
+                    label="Result of Execution",
+                    placeholder="Execution result will be shown here...",
+                    lines=2
+                )
+        
+        with gr.Row():
+            sample_btn = gr.Button("Example Prompt")
 
-            with gr.Row():
-                generate_btn = gr.Button("Generate", variant="primary")
-                clear_btn = gr.Button("Clear")
-                check_btn = gr.Button("Run test case")
+        with gr.Row():
+            generate_btn = gr.Button("Generate", variant="primary")
+            clear_btn = gr.Button("Clear")
+            check_btn = gr.Button("Run test case")
 
         # Generation Settings
         with gr.Accordion("Generation Settings"):
@@ -622,7 +624,7 @@ def create_chatbot_demo():
 
         sample_btn.click(
             fn=get_example_input,
-            outputs=[prefix_input, suffix_input, test_case_input, task_id_input],
+            outputs=[prefix_input, output_vis, suffix_input, test_case_input, task_id_input, result_output],
             queue=False
         )
     return demo

eval.py

@@ -1,204 +0,0 @@
-import ast
-import contextlib
-import faulthandler
-import io
-import multiprocessing
-import os
-import platform
-import signal
-import tempfile
-from typing import Callable, Dict, Optional
-
-
-import subprocess
-import time
-from multiprocessing import Process, Queue
-
-import io
-import sys
-
-def unsafe_execute(problem, completion, timeout=3):
-    check_program = (
-        problem["prompt"]
-        + completion
-        + problem["suffix"]
-        + "\n"
-        + problem["test"]
-        + "\n"
-        + f"check({problem['entry_point']})"
-    )
-
-    old_stdout = sys.stdout
-    old_stderr = sys.stderr
-    new_stdout = io.StringIO()
-    new_stderr = io.StringIO()
-    sys.stdout = new_stdout
-    sys.stderr = new_stderr
-
-    try:
-        exec(check_program, {})
-        output = new_stdout.getvalue().strip()
-        error_output = new_stderr.getvalue().strip()
-    except Exception as e:
-        output = ''
-        error_output = traceback.format_exc()
-    finally:
-        sys.stdout = old_stdout
-        sys.stderr = old_stderr
-
-    if output and not error_output:
-        return output
-    elif not output and not error_output:
-        return 'Pass all test cases!'
-    else:
-        return f'Error: {error_output}'
-
-
-@contextlib.contextmanager
-def time_limit(seconds: float):
-    def signal_handler(signum, frame):
-        raise TimeoutException("Timed out!")
-
-    signal.setitimer(signal.ITIMER_REAL, seconds)
-    signal.signal(signal.SIGALRM, signal_handler)
-    try:
-        yield
-    finally:
-        signal.setitimer(signal.ITIMER_REAL, 0)
-
-
-@contextlib.contextmanager
-def swallow_io():
-    stream = WriteOnlyStringIO()
-    with contextlib.redirect_stdout(stream):
-        with contextlib.redirect_stderr(stream):
-            with redirect_stdin(stream):
-                yield
-
-
-@contextlib.contextmanager
-def create_tempdir():
-    with tempfile.TemporaryDirectory() as dirname:
-        with chdir(dirname):
-            yield dirname
-
-
-class TimeoutException(Exception):
-    pass
-
-
-class WriteOnlyStringIO(io.StringIO):
-    """StringIO that throws an exception when it's read from"""
-
-    def read(self, *args, **kwargs):
-        raise IOError
-
-    def readline(self, *args, **kwargs):
-        raise IOError
-
-    def readlines(self, *args, **kwargs):
-        raise IOError
-
-    def readable(self, *args, **kwargs):
-        """Returns True if the IO object can be read."""
-        return False
-
-
-class redirect_stdin(contextlib._RedirectStream):  # type: ignore
-    _stream = "stdin"
-
-
-@contextlib.contextmanager
-def chdir(root):
-    if root == ".":
-        yield
-        return
-    cwd = os.getcwd()
-    os.chdir(root)
-    try:
-        yield
-    except BaseException as exc:
-        raise exc
-    finally:
-        os.chdir(cwd)
-
-
-def reliability_guard(maximum_memory_bytes: Optional[int] = None):
-    """
-    This disables various destructive functions and prevents the generated code
-    from interfering with the test (e.g. fork bomb, killing other processes,
-    removing filesystem files, etc.)
-
-    WARNING
-    This function is NOT a security sandbox. Untrusted code, including, model-
-    generated code, should not be blindly executed outside of one. See the
-    Codex paper for more information about OpenAI's code sandbox, and proceed
-    with caution.
-    """
-
-    if maximum_memory_bytes is not None:
-        import resource
-
-        resource.setrlimit(resource.RLIMIT_AS, (maximum_memory_bytes, maximum_memory_bytes))
-        resource.setrlimit(resource.RLIMIT_DATA, (maximum_memory_bytes, maximum_memory_bytes))
-        if not platform.uname().system == "Darwin":
-            resource.setrlimit(resource.RLIMIT_STACK, (maximum_memory_bytes, maximum_memory_bytes))
-
-    faulthandler.disable()
-
-    import builtins
-
-    builtins.exit = None
-    builtins.quit = None
-
-    import os
-
-    os.environ["OMP_NUM_THREADS"] = "1"
-
-    os.kill = None
-    os.system = None
-    os.putenv = None
-    os.remove = None
-    os.removedirs = None
-    os.rmdir = None
-    os.fchdir = None
-    os.setuid = None
-    os.fork = None
-    os.forkpty = None
-    os.killpg = None
-    os.rename = None
-    os.renames = None
-    os.truncate = None
-    os.replace = None
-    os.unlink = None
-    os.fchmod = None
-    os.fchown = None
-    os.chmod = None
-    os.chown = None
-    os.chroot = None
-    os.fchdir = None
-    os.lchflags = None
-    os.lchmod = None
-    os.lchown = None
-    os.getcwd = None
-    os.chdir = None
-
-    import shutil
-
-    shutil.rmtree = None
-    shutil.move = None
-    shutil.chown = None
-
-    import subprocess
-
-    subprocess.Popen = None  # type: ignore
-
-    __builtins__["help"] = None
-
-    import sys
-
-    sys.modules["ipdb"] = None
-    sys.modules["joblib"] = None
-    sys.modules["resource"] = None
-    sys.modules["psutil"] = None
-    sys.modules["tkinter"] = None