Upload 140 files

.gitattributes

@@ -33,3 +33,9 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+static/img/Accept_Reject_Reasons_Distribution.png filter=lfs diff=lfs merge=lfs -text
+static/img/Average_Ratings_Biased_Reviewers.png filter=lfs diff=lfs merge=lfs -text
+static/img/ICLR2025_uses_agents.png filter=lfs diff=lfs merge=lfs -text
+static/img/Overview.png filter=lfs diff=lfs merge=lfs -text
+static/img/ReviewPipeline.jpg filter=lfs diff=lfs merge=lfs -text
+static/img/ReviewPipeline.png filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -0,0 +1,185 @@
+
+
+key.py
+
+*.pdf
+*.json
+*.png
+*.jpg
+*.jpeg
+*.gif
+data/
+unused_data/
+demo/
+Summary/
+
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+outputs
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+pip-wheel-metadata/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+.python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+#poetry.lock
+
+# pdm
+#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
+#pdm.lock
+#   pdm stores project-wide configurations in .pdm.toml, but it is recommended to not include it
+#   in version control.
+#   https://pdm.fming.dev/#use-with-ide
+.pdm.toml
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+# PyCharm
+#  JetBrains specific template is maintained in a separate JetBrains.gitignore that can
+#  be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
+#  and can be added to the global gitignore or merged into this file.  For a more nuclear
+#  option (not recommended) you can uncomment the following to ignore the entire idea folder.
+.idea/
+
+.DS_Store
+hf-spaces/
+etc/
+.conda
+*.xlsx
+*.csv
+*.zip

LICENSE

@@ -0,0 +1,203 @@
+Copyright 2023 ChatArena. All rights reserved.
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

README.md

@@ -1,13 +1,18 @@
 ---
 title: AutoReview
-emoji: 👁
+emoji: 😊
 colorFrom: indigo
-colorTo: red
+colorTo: pink
 sdk: gradio
-sdk_version: 5.44.1
+sdk_version: 5.4.0
 app_file: app.py
 pinned: false
 license: apache-2.0
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+
+## License
+
+This project is licensed under the Apache-2.0 License.
+
+

app.py

@@ -0,0 +1,725 @@
+import json
+import re
+import os
+from datetime import datetime
+from glob import glob
+from argparse import Namespace
+
+import gradio as gr
+
+
+from autoreview import const
+from autoreview.config import AgentConfig
+from autoreview.agent import Player
+from autoreview.backends import BACKEND_REGISTRY
+from autoreview.environments import PaperReview
+from autoreview.paper_review_arena import PaperReviewArena
+from autoreview.utility.experiment_utils import initialize_players
+from autoreview.paper_review_player import PaperExtractorPlayer, Reviewer
+from autoreview.role_descriptions import (get_reviewer_description, get_paper_extractor_config)
+
+# 该文件的使命是前端交互：构建前端页面，从页面中获取用户的配置，传入后端运行，将结果实时展示到相应模块
+
+css = """
+/* 全局样式 */
+* {
+    font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif;
+}
+
+/* 主容器 */
+#col-container {
+    max-width: 95%; 
+    margin: 0 auto; 
+    display: flex; 
+    flex-direction: column;
+    background: #e3f2fd;
+    min-height: 100vh;
+    padding: 20px;
+    border-radius: 15px;
+    box-shadow: 0 20px 40px rgba(0,0,0,0.1);
+}
+
+/* 头部样式 */
+#header {
+    text-align: center;
+    background: rgba(255,255,255,0.95);
+    padding: 30px;
+    border-radius: 15px;
+    margin-bottom: 20px;
+    box-shadow: 0 10px 30px rgba(0,0,0,0.1);
+}
+
+#header h1 {
+    color: #2c3e50;
+    font-size: 2.5em;
+    margin: 0 0 10px 0;
+    font-weight: 700;
+}
+
+#header p {
+    color: #7f8c8d;
+    font-size: 1.1em;
+    margin: 0;
+}
+
+/* 聊天区域 */
+#col-chatbox {
+    flex: 1; 
+    max-height: min(900px, 100%);
+    background: rgba(255,255,255,0.95);
+    border-radius: 15px;
+    padding: 20px;
+    box-shadow: 0 10px 30px rgba(0,0,0,0.1);
+}
+
+
+/* 配置区域 */
+#col-config {
+    background: rgba(255,255,255,0.95);
+    border-radius: 15px;
+    padding: 20px;
+    box-shadow: 0 10px 30px rgba(0,0,0,0.1);
+    margin-left: 20px;
+}
+
+/* 标签样式 */
+#label {
+    font-size: 2em; 
+    padding: 0.5em; 
+    margin: 0;
+    color: #2c3e50;
+}
+
+/* 消息样式 */
+.message {
+    font-size: 1.1em;
+    line-height: 1.6;
+}
+
+.message-wrap {
+    max-height: min(700px, 100vh);
+}
+
+/* 按钮样式 */
+button {
+    background: #FFFFFF;
+    border: none;
+    color: black;
+    padding: 12px 24px;
+    border-radius: 25px;
+    font-weight: 600;
+    transition: all 0.3s ease;
+}
+
+button:hover {
+    transform: translateY(-2px);
+}
+
+/* 下拉框样式 */
+select, .gr-dropdown {
+    border-radius: 10px;
+    border: 2px solid #e0e0e0;
+    padding: 8px 12px;
+    transition: all 0.3s ease;
+}
+
+select:focus, .gr-dropdown:focus {
+    border-color: #FFFFFF;
+    box-shadow: 0 0 0 3px rgba(100, 181, 246, 0.1);
+}
+
+/* 文本框样式 */
+textarea, .gr-textbox {
+    border-radius: 10px;
+    border: 2px solid #e0e0e0;
+    padding: 12px;
+    transition: all 0.3s ease;
+}
+
+textarea:focus, .gr-textbox:focus {
+    border-color: #FFFFFF;
+    box-shadow: 0 0 0 3px rgba(100, 181, 246, 0.1);
+}
+
+/* 滑块样式 */
+.gr-slider {
+    border-radius: 10px;
+}
+
+/* 标签页样式 */
+.tabs {
+    background: rgba(255,255,255,0.9);
+    border-radius: 10px;
+    padding: 15px;
+    margin-bottom: 15px;
+}
+
+/* 文件上传区域 */
+.gr-file, .file-upload, .file-upload-area, .file-drop-zone {
+    border: 2px dashed #FFFFFF;
+    border-radius: 15px;
+    padding: 20px;
+    text-align: center;
+    background: rgba(100, 181, 246, 0.05);
+    transition: all 0.3s ease;
+    color: white !important;
+}
+
+.gr-file:hover, .file-upload:hover, .file-upload-area:hover, .file-drop-zone:hover {
+    background: rgba(100, 181, 246, 0.1);
+    border-color: #42a5f5;
+    color: white !important;
+}
+
+/* 文件上传组件内的文本 */
+.gr-file *, .file-upload *, .file-upload-area *, .file-drop-zone * {
+    color: white !important;
+}
+
+/* 响应式设计 */
+@media (max-width: 768px) {
+    #col-container {
+        max-width: 100%;
+        padding: 10px;
+    }
+    
+    #col-config {
+        margin-left: 0;
+        margin-top: 20px;
+    }
+}
+"""
+# .wrap {min-width: min(640px, 100vh)}
+# #env-desc {max-height: 100px; overflow-y: auto;}
+# .textarea {height: 100px; max-height: 100px;}
+# #chatbot-tab-all {height: 750px; max-height: min(750px, 100%);}
+# #chatbox {height: min(750px, 100%); max-height: min(750px, 100%);}
+# #chatbox.block {height: 730px}
+# .wrap {max-height: 680px;}
+# .scroll-hide {overflow-y: scroll; max-height: 100px;}
+
+DEBUG = False
+
+DEFAULT_BACKEND = "openai-chat"
+MAX_NUM_PLAYERS = 3  # Changed from 5 to 3 - only reviewers
+DEFAULT_NUM_PLAYERS = 3  # Changed from 5 to 3 - only reviewers
+CURRENT_STEP_INDEX = 0
+
+os.environ['OPENAI_API_KEY'] = 'sk-Aors1iVXAbgd7sGwC9Ff781c75D14b74A71d4e63F1E46b68'
+os.environ['OPENAI_BASEURL'] = 'https://api2.aigcbest.top/v1'
+
+
+
+def load_examples():
+    example_configs = {}
+    # Load json config files from examples folder
+    example_files = glob("examples/*.json")
+    for example_file in example_files:
+        with open(example_file, encoding="utf-8") as f:
+            example = json.load(f)
+            try:
+                example_configs[example["name"]] = example
+            except KeyError:
+                print(f"Example {example_file} is missing a name field. Skipping.")
+    return example_configs
+
+
+EXAMPLE_REGISTRY = load_examples()
+
+# DB = SupabaseDB() if supabase_available else None
+
+def get_player_components(name, visible):
+    with gr.Row():
+        with gr.Column():
+            role_name = gr.Textbox(
+                lines=1,
+                show_label=False,
+                interactive=True,
+                visible=False,
+                value=name,
+            )
+            
+            # is benign, is_knowledgeable, is_responsible, 
+            # player_config = gr.CheckboxGroup(
+            #     choices=["Benign", "Knowledgeable", "Responsible"],
+            #     label="Reviewer Type",
+            #     visible=visible,
+            # )
+            
+            gr.Markdown(f"### 👨‍💼 {name} Settings")
+            with gr.Row():
+                Expertise_config = gr.Dropdown(
+                    choices=["Machine Learning", "Natural Language Processing", "Computer Vision", "Robotics", "Data Science", "Software Engineering", 
+                            "Bioinformatics", "Computational Biology", "Genomics", "Proteomics", "Systems Biology", "Synthetic Biology", 
+                            "Biomedical Engineering", "Neuroscience", "Drug Discovery", "Medical Imaging", "General"],
+                    interactive=True,
+                    label = "🎯 Expertise Area",
+                    show_label=True,
+                    value="General",
+                    allow_custom_value=True
+                )
+                paper_type_box = gr.Dropdown(
+                        choices=["Technical Paper", "Survey Paper", "Application Paper", "Dataset Paper", "Tool Paper"],
+                        interactive=True,
+                        label="📝 Paper Type",
+                        show_label=True,
+                        value="Technical Paper",
+                    )
+                
+                
+            
+            role_desc = gr.Textbox(
+                lines=8,
+                max_lines=8,
+                show_label=True,
+                label="📋 Reviewer Description",
+                interactive=True,
+                visible=visible,
+                autoscroll=False,
+                value=get_reviewer_description()
+            )
+
+            def update_role_desc(Expertise_config, paper_type):
+                
+                # Always use reviewer_write_reviews phase since we're only doing Phase I
+                phase = 'reviewer_write_reviews'
+                return get_reviewer_description(expertise=Expertise_config, paper_type=paper_type, phase=phase)
+                
+            Expertise_config.select(fn=update_role_desc, inputs=[Expertise_config, paper_type_box], outputs=[role_desc])
+            paper_type_box.select(fn=update_role_desc, inputs=[Expertise_config, paper_type_box], outputs=[role_desc])
+            
+        with gr.Column():
+            with gr.Accordion(
+                f"⚙️ {name} Configuration", open=True, visible=visible
+            ) as accordion:
+                temperature = gr.Slider(
+                    minimum=0.,
+                    maximum=2.0,
+                    step=0.1,
+                    interactive=True,
+                    visible=visible,
+                    label="🌡️ Temperature (Creativity)",
+                    value=1.0,
+                )
+                max_tokens = gr.Slider(
+                    minimum=10,
+                    maximum=600,
+                    step=10,
+                    interactive=True,
+                    visible=visible,
+                    label="📏 Max Tokens",
+                    value=600,
+                )
+
+    return [role_name, Expertise_config, paper_type_box, accordion, temperature, max_tokens]
+
+
+
+
+
+def get_empty_state():
+    return gr.State({"arena": None})
+
+
+with (gr.Blocks(css=css) as demo):
+    state = get_empty_state()
+    all_components = []
+
+    with gr.Column(elem_id="col-container"):
+        gr.Markdown(
+            """
+# 🤖 AI Paper Review System
+## Intelligent Manuscript Evaluation with Multi-Agent Reviewers
+
+**Transform your paper review process with AI-powered reviewers that provide comprehensive feedback.**
+
+---
+""",
+            elem_id="header",
+        )
+        
+
+        
+        with gr.Row():
+            with gr.Column(elem_id="col-chatbox"):
+                gr.Markdown("### 💬 Review Progress")
+
+                player_chatbots = []
+                for i in range(MAX_NUM_PLAYERS):
+                    player_name = f"Reviewer {i + 1}"
+
+                    with gr.Tab(f"👨‍💼 {player_name}", visible=(i < DEFAULT_NUM_PLAYERS)):
+                        player_chatbot = gr.Chatbot(
+                            elem_id=f"chatbox-{i}",
+                            visible=i < DEFAULT_NUM_PLAYERS,
+                            label=player_name,
+                            show_label=False,
+                            height=600  # FIXME: this parameter is not working
+                        )
+                        player_chatbots.append(player_chatbot)
+
+            all_components += [*player_chatbots]
+
+            with gr.Column(elem_id="col-config"):
+                gr.Markdown("### ⚙️ Configuration Panel")
+                # gr.Markdown("Player Configuration")
+                
+                # parallel_checkbox = gr.Checkbox(
+                #     label="Parallel Actions", value=False, visible=True
+                # )
+                
+                all_players_components, players_idx2comp = [], {}
+                with gr.Blocks():
+                    for i in range(MAX_NUM_PLAYERS):
+                        player_name = f"Reviewer {i + 1}"
+
+                        with gr.Tab(
+                            f"👨‍💼 {player_name}", visible=(i < DEFAULT_NUM_PLAYERS)
+                        ) as tab:
+                            player_comps = get_player_components(
+                                player_name, visible=(i < DEFAULT_NUM_PLAYERS)
+                            )
+
+                        players_idx2comp[i] = player_comps + [tab]
+                        all_players_components += player_comps + [tab]
+
+                
+                all_components += all_players_components
+                
+                
+                gr.Markdown("### 📄 Upload Your Paper")
+                upload_file_box = gr.File(
+                    visible=True,
+                    height=120,
+                    label="Choose PDF file",
+                    file_types=[".pdf"],
+                    file_count="single"
+                )
+                
+                with gr.Row():
+                    btn_step = gr.Button("🚀 Start Review", variant="primary", size="lg")
+                    btn_restart = gr.Button("🔄 Reset", variant="secondary", size="lg")
+
+                all_components += [upload_file_box, btn_step, btn_restart]
+    
+    
+    def _convert_to_chatbot_output(all_messages, display_recv=False):
+        chatbot_output = []
+        try:
+            for i, message in enumerate(all_messages):
+                # Handle case where message might be None or invalid
+                if not hasattr(message, 'agent_name') or not hasattr(message, 'content'):
+                    print(f"Warning: Invalid message at index {i}: {message}")
+                    continue
+                    
+                agent_name, msg, recv = (
+                    message.agent_name,
+                    message.content,
+                    str(message.visible_to),
+                )
+                
+                # Handle empty or None content
+                if not msg or msg.strip() == "":
+                    print(f"Warning: Empty message from {agent_name}")
+                    continue
+                    
+                new_msg = re.sub(
+                    r"\n+", "<br>", msg.strip()
+                )  # Preprocess message for chatbot output
+                if display_recv:
+                    new_msg = f"**{agent_name} (-> {recv})**: {new_msg}"  # Add role to the message
+                else:
+                    new_msg = f"**{agent_name}**: {new_msg}"
+
+                # Convert to Gradio Chatbot format: list of [user_message, bot_message] tuples
+                if agent_name == "Moderator":
+                    chatbot_output.append([new_msg, None])
+                else:
+                    chatbot_output.append([None, new_msg])
+        except Exception as e:
+            print(f"Error in _convert_to_chatbot_output: {e}")
+            print(f"all_messages type: {type(all_messages)}")
+            print(f"all_messages content: {all_messages}")
+            # Return empty list as fallback
+            return []
+        return chatbot_output
+    
+    def _create_arena_config_from_components(all_comps: dict):
+        
+        env_desc = const.GLOBAL_PROMPT
+        paper_pdf_path = all_comps[upload_file_box]
+        
+        # Step 1: Initialize the players
+        num_players = MAX_NUM_PLAYERS
+        
+        # You can ignore these fields for the demo
+        conference = "EMNLP2024"
+        paper_decision = "Accept"
+        data_dir = ''
+        paper_id = "12345"
+
+        args = Namespace(openai_client_type="openai",
+                         experiment_name="test",
+                         max_num_words=16384)
+        
+        # Phase I: Only reviewers are active
+        players = []
+        
+        # 不能直接获取role_desc，需要根据Intention_config, Knowledge_config, Responsibility_config生成一个配置
+        # self.environment.experiment_setting["players"]['Reviewer'][reviewer_index - 1]
+        
+        experiment_setting = {
+        "paper_id": paper_id,
+        "paper_decision": paper_decision,
+        "players": {
+
+            # Paper Extractor is a special player that extracts a paper from the dataset.
+            # Its constructor does not take any arguments.
+            "Paper Extractor": [{}],
+
+            # Reviewer settings are generated based on reviewer types provided in the settings.
+            "Reviewer": [],
+        },
+            # "global_settings": setting['global_settings']
+        }
+        
+        
+        for i in range(num_players):
+
+            role_name = role_desc = temperature = max_tokens = None
+
+            # All players are reviewers in Phase I only
+            role_name, expertise_config, paper_type, temperature, max_tokens = (
+                all_comps[c]
+                for c in players_idx2comp[i]
+                if not isinstance(c, (gr.Accordion, gr.Tab))
+            )
+            
+            # Use default backend
+            backend_type = DEFAULT_BACKEND
+            
+            # Use default values for other attributes since we removed them from UI
+            experiment_setting["players"]['Reviewer'].append({"is_benign": None, 
+                                                                "is_knowledgeable": None, 
+                                                                "is_responsible": None,
+                                                                "knows_authors": 'unfamous'})
+                
+            role_desc = get_reviewer_description(expertise=expertise_config, paper_type=paper_type)
+
+
+
+            # common config for all players
+            player_config = {
+                "name": role_name,
+                "role_desc": role_desc,
+                "global_prompt": env_desc,
+                "backend": {
+                    "backend_type": backend_type,
+                    "temperature": temperature,
+                    "max_tokens": max_tokens,
+                },
+            }
+            
+            player_config = AgentConfig(**player_config)
+            
+            # All players are reviewers in Phase I only
+            player = Reviewer(data_dir=data_dir, conference=conference, args=args, **player_config)
+            
+            players.append(player)
+            
+        # Phase I: Create paper extractor for reviewer assessment
+            
+        # Add paper extractor
+        paper_extractor_config = get_paper_extractor_config(max_tokens=2048)
+
+        paper_extractor = PaperExtractorPlayer(paper_pdf_path=paper_pdf_path,
+                                            data_dir=data_dir, paper_id=paper_id,
+                                            paper_decision=paper_decision, args=args,
+                                            conference=conference, **paper_extractor_config)
+        players.append(paper_extractor)
+            
+        
+        player_names = [player.name for player in players]
+        
+        # Step 2: Initialize the environment
+        env = PaperReview(player_names=player_names, paper_decision=paper_decision, paper_id=paper_id,
+                          args=args, experiment_setting=experiment_setting)
+        
+        # Step 3: Initialize the Arena
+        arena = PaperReviewArena(players=players, environment=env, args=args, global_prompt=env_desc)
+        
+        return arena
+        
+    def step_game(all_comps: dict):
+        global CURRENT_STEP_INDEX
+
+        yield {
+            btn_step: gr.update(value="Running...", interactive=False),
+            btn_restart: gr.update(interactive=False),
+        }
+        
+        cur_state = all_comps[state]
+        
+        # If arena is not yet created, create it
+        if cur_state["arena"] is None:
+            # Create the Arena
+            arena = _create_arena_config_from_components(all_comps)
+            cur_state["arena"] = arena
+        else:
+            arena = cur_state["arena"]
+        
+        # TODO: 连续运行
+        
+        timestep = arena.step()
+        
+        CURRENT_STEP_INDEX = int(arena.environment.phase_index)
+        
+        # 更新前端信息
+        if timestep:
+            try:
+                all_messages = timestep.observation
+
+                # Initialize update dictionary
+                update_dict = {
+                    btn_step: gr.update(
+                        value="Next Step", interactive=not timestep.terminal
+                    ),
+                    btn_restart: gr.update(interactive=True),
+                    state: cur_state,
+                }
+            except Exception as e:
+                print(f"Error updating main chatbot: {e}")
+                # Initialize update dictionary with empty chatbot output
+                update_dict = {
+                    btn_step: gr.update(
+                        value="Next Step", interactive=not timestep.terminal
+                    ),
+                    btn_restart: gr.update(interactive=True),
+                    state: cur_state,
+                }
+
+            # Define a mapping of player names to their respective chatbots
+            player_name_to_chatbot = {
+                "Reviewer 1": player_chatbots[0],
+                "Reviewer 2": player_chatbots[1],
+                "Reviewer 3": player_chatbots[2]
+            }
+
+            # Update each player's chatbot output
+            for player in arena.players:
+                player_name = player.name
+                if player_name in player_name_to_chatbot:
+                    try:
+                        player_messages = arena.environment.get_messages_from_player(player_name)
+                        # player_messages[0].content = 'Paper content has been extracted.'
+                        player_output = _convert_to_chatbot_output(player_messages)
+                        update_dict[player_name_to_chatbot[player_name]] = player_output
+                    except Exception as e:
+                        print(f"Error updating player {player_name} chatbot: {e}")
+                        # Set empty output as fallback
+                        update_dict[player_name_to_chatbot[player_name]] = []
+
+            # # Reviewer 1, 2, 3 Area Chair, Paper Extractor, Author
+            # for i, player in enumerate(arena.players):
+            #     player_name = player.name
+            #     # Get the messages for the current player
+            #     player_messages = arena.environment.get_observation(player_name)
+            #     player_messages[0].content = 'Paper content has been extracted.'
+            #
+            #     # Convert messages to chatbot output
+            #     player_output = _convert_to_chatbot_output(player_messages)
+
+
+                """
+                # Phase I only: All players are reviewers
+                if 'Reviewer' in player.name:
+                    player_messages = arena.environment.get_observation(player.name)
+                    # 不要显示第一条长段的信息，只显示 文章内容已被抽取
+                    player_messages[0].content = 'Paper content has been extracted.'
+                    player_output = _convert_to_chatbot_output(player_messages)
+                    # Update the player's chatbot output
+                    update_dict[player_chatbots[i]] = player_output
+                """
+            # Ahren: Auto run
+            # if not timestep.terminal:
+            #     yield from step_game(all_comps)
+
+            yield update_dict
+        
+            
+    def restart_game(all_comps: dict):
+        global CURRENT_STEP_INDEX
+        CURRENT_STEP_INDEX = 0
+        
+        cur_state = all_comps[state]
+        cur_state["arena"] = None
+        yield {
+            btn_restart: gr.update(interactive=False),
+            btn_step: gr.update(interactive=False),
+            state: cur_state,
+        }
+
+        # arena_config = _create_arena_config_from_components(all_comps)
+        # arena = Arena.from_config(arena_config)
+        # log_arena(arena, database=DB)
+        # cur_state["arena"] = arena
+
+        yield {
+            btn_step: gr.update(value="Start", interactive=True),
+            btn_restart: gr.update(interactive=True),
+            upload_file_box: gr.update(value=None),
+            state: cur_state,
+        }
+    
+    # Remove Accordion and Tab from the list of components
+    all_components = [
+        comp for comp in all_components if not isinstance(comp, (gr.Accordion, gr.Tab))
+    ]
+    
+    # update component
+
+    # If any of the Textbox, Slider, Checkbox, Dropdown, RadioButtons is changed, the Step button is disabled
+    for comp in all_components:
+
+        def _disable_step_button(state):
+            if state["arena"] is not None:
+                return gr.update(interactive=False)
+            else:
+                return gr.update()
+
+        if (
+            isinstance(
+                comp, (gr.Textbox, gr.Slider, gr.Checkbox, gr.Dropdown, gr.Radio)
+            )
+            and comp is not upload_file_box
+        ):
+            comp.change(_disable_step_button, state, btn_step)
+
+    # Add paper type change event to update all reviewer descriptions
+    def update_all_reviewer_descriptions(paper_type, *reviewer_comps):
+        # This function will be called when paper type changes
+        # For now, we'll just disable the step button
+        return gr.update(interactive=False)
+    
+    # paper_type_box.change(
+    #     fn=update_all_reviewer_descriptions,
+    #     inputs=[paper_type_box] + [comp for comp in all_components if "Expertise_config" in str(comp)],
+    #     outputs=[btn_step]
+    # )
+
+    # Ahren: Auto run
+    btn_step.click(
+        step_game,
+        set(all_components + [state]),
+        [*player_chatbots, btn_step, btn_restart, state, upload_file_box],
+    )
+    
+    btn_restart.click(
+        restart_game,
+        set(all_components + [state]),
+        [*player_chatbots, btn_step, btn_restart, state, upload_file_box],
+    )
+
+    
+demo.queue()
+demo.launch()

app_v1.py

@@ -0,0 +1,807 @@
+import json
+import re
+import os
+from datetime import datetime
+from glob import glob
+from argparse import Namespace
+
+import gradio as gr
+
+
+from autoreview import const
+from autoreview.config import AgentConfig
+from autoreview.agent import Player
+from autoreview.backends import BACKEND_REGISTRY
+from autoreview.environments import PaperReview
+from autoreview.paper_review_arena import PaperReviewArena
+from autoreview.utility.experiment_utils import initialize_players
+from autoreview.paper_review_player import PaperExtractorPlayer, Reviewer
+from autoreview.role_descriptions import (get_reviewer_description, get_paper_extractor_config)
+from autoreview.paper_type_classifier import PaperTypeClassifier
+from llama_index.readers.file.docs import PDFReader
+
+# 该文件的使命是前端交互：构建前端页面，从页面中获取用户的配置，传入后端运行，将结果实时展示到相应模块
+
+css = """
+/* 全局样式 */
+* {
+    font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif;
+}
+
+/* 主容器 */
+#col-container {
+    max-width: 95%; 
+    margin: 0 auto; 
+    display: flex; 
+    flex-direction: column;
+    background: #e3f2fd;
+    min-height: 100vh;
+    padding: 20px;
+    border-radius: 15px;
+    box-shadow: 0 20px 40px rgba(0,0,0,0.1);
+}
+
+/* 头部样式 */
+#header {
+    text-align: center;
+    background: rgba(255,255,255,0.95);
+    padding: 30px;
+    border-radius: 15px;
+    margin-bottom: 20px;
+    box-shadow: 0 10px 30px rgba(0,0,0,0.1);
+}
+
+#header h1 {
+    color: #2c3e50;
+    font-size: 2.5em;
+    margin: 0 0 10px 0;
+    font-weight: 700;
+}
+
+#header p {
+    color: #7f8c8d;
+    font-size: 1.1em;
+    margin: 0;
+}
+
+/* 聊天区域 */
+#col-chatbox {
+    flex: 1; 
+    max-height: min(900px, 100%);
+    background: rgba(255,255,255,0.95);
+    border-radius: 15px;
+    padding: 20px;
+    box-shadow: 0 10px 30px rgba(0,0,0,0.1);
+}
+
+
+/* 配置区域 */
+#col-config {
+    background: rgba(255,255,255,0.95);
+    border-radius: 15px;
+    padding: 20px;
+    box-shadow: 0 10px 30px rgba(0,0,0,0.1);
+    margin-left: 20px;
+}
+
+/* 标签样式 */
+#label {
+    font-size: 2em; 
+    padding: 0.5em; 
+    margin: 0;
+    color: #2c3e50;
+}
+
+/* 消息样式 */
+.message {
+    font-size: 1.1em;
+    line-height: 1.6;
+}
+
+.message-wrap {
+    max-height: min(700px, 100vh);
+}
+
+/* 按钮样式 */
+button {
+    background: #FFFFFF;
+    border: none;
+    color: black;
+    padding: 12px 24px;
+    border-radius: 25px;
+    font-weight: 600;
+    transition: all 0.3s ease;
+}
+
+button:hover {
+    transform: translateY(-2px);
+}
+
+/* 下拉框样式 */
+select, .gr-dropdown {
+    border-radius: 10px;
+    border: 2px solid #e0e0e0;
+    padding: 8px 12px;
+    transition: all 0.3s ease;
+}
+
+select:focus, .gr-dropdown:focus {
+    border-color: #FFFFFF;
+    box-shadow: 0 0 0 3px rgba(100, 181, 246, 0.1);
+}
+
+/* 文本框样式 */
+textarea, .gr-textbox {
+    border-radius: 10px;
+    border: 2px solid #e0e0e0;
+    padding: 12px;
+    transition: all 0.3s ease;
+}
+
+textarea:focus, .gr-textbox:focus {
+    border-color: #FFFFFF;
+    box-shadow: 0 0 0 3px rgba(100, 181, 246, 0.1);
+}
+
+/* 滑块样式 */
+.gr-slider {
+    border-radius: 10px;
+}
+
+/* 标签页样式 */
+.tabs {
+    background: rgba(255,255,255,0.9);
+    border-radius: 10px;
+    padding: 15px;
+    margin-bottom: 15px;
+}
+
+/* 文件上传区域 */
+.gr-file, .file-upload, .file-upload-area, .file-drop-zone {
+    border: 2px dashed #FFFFFF;
+    border-radius: 15px;
+    padding: 20px;
+    text-align: center;
+    background: rgba(100, 181, 246, 0.05);
+    transition: all 0.3s ease;
+    color: white !important;
+}
+
+.gr-file:hover, .file-upload:hover, .file-upload-area:hover, .file-drop-zone:hover {
+    background: rgba(100, 181, 246, 0.1);
+    border-color: #42a5f5;
+    color: white !important;
+}
+
+/* 文件上传组件内的文本 */
+.gr-file *, .file-upload *, .file-upload-area *, .file-drop-zone * {
+    color: white !important;
+}
+
+/* 响应式设计 */
+@media (max-width: 768px) {
+    #col-container {
+        max-width: 100%;
+        padding: 10px;
+    }
+    
+    #col-config {
+        margin-left: 0;
+        margin-top: 20px;
+    }
+}
+"""
+# .wrap {min-width: min(640px, 100vh)}
+# #env-desc {max-height: 100px; overflow-y: auto;}
+# .textarea {height: 100px; max-height: 100px;}
+# #chatbot-tab-all {height: 750px; max-height: min(750px, 100%);}
+# #chatbox {height: min(750px, 100%); max-height: min(750px, 100%);}
+# #chatbox.block {height: 730px}
+# .wrap {max-height: 680px;}
+# .scroll-hide {overflow-y: scroll; max-height: 100px;}
+
+DEBUG = False
+
+DEFAULT_BACKEND = "openai-chat"
+MAX_NUM_PLAYERS = 3  # Changed from 5 to 3 - only reviewers
+DEFAULT_NUM_PLAYERS = 3  # Changed from 5 to 3 - only reviewers
+CURRENT_STEP_INDEX = 0
+
+os.environ['OPENAI_API_KEY'] = 'sk-Aors1iVXAbgd7sGwC9Ff781c75D14b74A71d4e63F1E46b68'
+os.environ['OPENAI_BASEURL'] = 'https://api2.aigcbest.top/v1'
+
+
+
+def load_examples():
+    example_configs = {}
+    # Load json config files from examples folder
+    example_files = glob("examples/*.json")
+    for example_file in example_files:
+        with open(example_file, encoding="utf-8") as f:
+            example = json.load(f)
+            try:
+                example_configs[example["name"]] = example
+            except KeyError:
+                print(f"Example {example_file} is missing a name field. Skipping.")
+    return example_configs
+
+
+EXAMPLE_REGISTRY = load_examples()
+
+# DB = SupabaseDB() if supabase_available else None
+
+def auto_detect_paper_type(pdf_file):
+    """Automatically detect paper type from uploaded PDF file."""
+    if pdf_file is None:
+        return "Technical Paper"  # Default fallback
+    
+    try:
+        # Create a temporary paper extractor to get paper content
+        from pathlib import Path
+        import tempfile
+        
+        # Extract text from PDF
+        loader = PDFReader()
+        document_path = Path(pdf_file.name)
+        documents = loader.load_data(file=document_path)
+        
+        # Combine all text
+        paper_content = ""
+        for doc in documents:
+            paper_content += doc.text + " "
+        
+        # Use classifier to detect paper type
+        classifier = PaperTypeClassifier()
+        detected_type = classifier.classify_paper(paper_content[:2000], use_llm=True)
+        
+        return detected_type
+        
+    except Exception as e:
+        print(f"Error in auto-detecting paper type: {e}")
+        return "Technical Paper"  # Default fallback
+
+def update_all_paper_types(detected_type, *paper_type_boxes):
+    """Update all paper type dropdowns with the detected type."""
+    # Extract the actual type from the status text
+    if isinstance(detected_type, str) and "**Paper Type:**" in detected_type:
+        # Extract type from status text like "**Paper Type:** Technical Paper (auto-detected)"
+        type_part = detected_type.split("**Paper Type:**")[1].split("(")[0].strip()
+    else:
+        type_part = detected_type
+    
+    return [type_part] * len(paper_type_boxes)
+
+def update_paper_type_status(detected_type):
+    """Update the paper type status display."""
+    return f"**Paper Type:** {detected_type} (auto-detected)"
+
+def get_player_components(name, visible):
+    with gr.Row():
+        with gr.Column():
+            role_name = gr.Textbox(
+                lines=1,
+                show_label=False,
+                interactive=True,
+                visible=False,
+                value=name,
+            )
+            
+            # is benign, is_knowledgeable, is_responsible, 
+            # player_config = gr.CheckboxGroup(
+            #     choices=["Benign", "Knowledgeable", "Responsible"],
+            #     label="Reviewer Type",
+            #     visible=visible,
+            # )
+            
+            gr.Markdown(f"### 👨‍💼 {name} Settings")
+            with gr.Row():
+                Expertise_config = gr.Dropdown(
+                    choices=["Machine Learning", "Natural Language Processing", "Computer Vision", "Robotics", "Data Science", "Software Engineering", 
+                            "Bioinformatics", "Computational Biology", "Genomics", "Proteomics", "Systems Biology", "Synthetic Biology", 
+                            "Biomedical Engineering", "Neuroscience", "Drug Discovery", "Medical Imaging", "General"],
+                    interactive=True,
+                    label = "🎯 Expertise Area",
+                    show_label=True,
+                    value="General",
+                    allow_custom_value=True
+                )
+                paper_type_box = gr.Dropdown(
+                        choices=["Technical Paper", "Survey Paper", "Application Paper", "Dataset Paper", "Tool Paper"],
+                        interactive=True,
+                        label="📝 Paper Type",
+                        show_label=True,
+                        value="Technical Paper",
+                    )
+                
+                
+            
+            role_desc = gr.Textbox(
+                lines=8,
+                max_lines=8,
+                show_label=True,
+                label="📋 Reviewer Description",
+                interactive=True,
+                visible=visible,
+                autoscroll=False,
+                value=get_reviewer_description()
+            )
+
+            def update_role_desc(Expertise_config, paper_type):
+                
+                # Always use reviewer_write_reviews phase since we're only doing Phase I
+                phase = 'reviewer_write_reviews'
+                return get_reviewer_description(expertise=Expertise_config, paper_type=paper_type, phase=phase)
+                
+            Expertise_config.select(fn=update_role_desc, inputs=[Expertise_config, paper_type_box], outputs=[role_desc])
+            paper_type_box.select(fn=update_role_desc, inputs=[Expertise_config, paper_type_box], outputs=[role_desc])
+            
+        with gr.Column():
+            with gr.Accordion(
+                f"⚙️ {name} Configuration", open=True, visible=visible
+            ) as accordion:
+                temperature = gr.Slider(
+                    minimum=0.,
+                    maximum=2.0,
+                    step=0.1,
+                    interactive=True,
+                    visible=visible,
+                    label="🌡️ Temperature (Creativity)",
+                    value=1.0,
+                )
+                max_tokens = gr.Slider(
+                    minimum=10,
+                    maximum=600,
+                    step=10,
+                    interactive=True,
+                    visible=visible,
+                    label="📏 Max Tokens",
+                    value=600,
+                )
+
+    return [role_name, Expertise_config, paper_type_box, accordion, temperature, max_tokens]
+
+
+
+
+
+def get_empty_state():
+    return gr.State({"arena": None})
+
+
+with (gr.Blocks(css=css) as demo):
+    state = get_empty_state()
+    all_components = []
+
+    with gr.Column(elem_id="col-container"):
+        gr.Markdown(
+            """
+# 🤖 AI Paper Review System
+## Intelligent Manuscript Evaluation with Multi-Agent Reviewers
+
+**Transform your paper review process with AI-powered reviewers that provide comprehensive feedback.**
+
+---
+""",
+            elem_id="header",
+        )
+        
+
+        
+        with gr.Row():
+            with gr.Column(elem_id="col-chatbox"):
+                gr.Markdown("### 💬 Review Progress")
+
+                player_chatbots = []
+                for i in range(MAX_NUM_PLAYERS):
+                    player_name = f"Reviewer {i + 1}"
+
+                    with gr.Tab(f"👨‍💼 {player_name}", visible=(i < DEFAULT_NUM_PLAYERS)):
+                        player_chatbot = gr.Chatbot(
+                            elem_id=f"chatbox-{i}",
+                            visible=i < DEFAULT_NUM_PLAYERS,
+                            label=player_name,
+                            show_label=False,
+                            height=600  # FIXME: this parameter is not working
+                        )
+                        player_chatbots.append(player_chatbot)
+
+            all_components += [*player_chatbots]
+
+            with gr.Column(elem_id="col-config"):
+                gr.Markdown("### ⚙️ Configuration Panel")
+                # gr.Markdown("Player Configuration")
+                
+                # parallel_checkbox = gr.Checkbox(
+                #     label="Parallel Actions", value=False, visible=True
+                # )
+                
+                all_players_components, players_idx2comp = [], {}
+                with gr.Blocks():
+                    for i in range(MAX_NUM_PLAYERS):
+                        player_name = f"Reviewer {i + 1}"
+
+                        with gr.Tab(
+                            f"👨‍💼 {player_name}", visible=(i < DEFAULT_NUM_PLAYERS)
+                        ) as tab:
+                            player_comps = get_player_components(
+                                player_name, visible=(i < DEFAULT_NUM_PLAYERS)
+                            )
+
+                        players_idx2comp[i] = player_comps + [tab]
+                        all_players_components += player_comps + [tab]
+
+                
+                all_components += all_players_components
+                
+                
+                gr.Markdown("### 📄 Upload Your Paper")
+                upload_file_box = gr.File(
+                    visible=True,
+                    height=120,
+                    label="Choose PDF file",
+                    file_types=[".pdf"],
+                    file_count="single"
+                )
+                
+                # Add auto-detect button
+                with gr.Row():
+                    auto_detect_btn = gr.Button("🔍 Auto-detect Paper Type", variant="secondary", size="sm")
+                    paper_type_status = gr.Markdown("**Paper Type:** Technical Paper (default)")
+                
+                with gr.Row():
+                    btn_step = gr.Button("🚀 Start Review", variant="primary", size="lg")
+                    btn_restart = gr.Button("🔄 Reset", variant="secondary", size="lg")
+
+                all_components += [upload_file_box, auto_detect_btn, paper_type_status, btn_step, btn_restart]
+    
+    
+    def _convert_to_chatbot_output(all_messages, display_recv=False):
+        chatbot_output = []
+        try:
+            for i, message in enumerate(all_messages):
+                # Handle case where message might be None or invalid
+                if not hasattr(message, 'agent_name') or not hasattr(message, 'content'):
+                    print(f"Warning: Invalid message at index {i}: {message}")
+                    continue
+                    
+                agent_name, msg, recv = (
+                    message.agent_name,
+                    message.content,
+                    str(message.visible_to),
+                )
+                
+                # Handle empty or None content
+                if not msg or msg.strip() == "":
+                    print(f"Warning: Empty message from {agent_name}")
+                    continue
+                    
+                new_msg = re.sub(
+                    r"\n+", "<br>", msg.strip()
+                )  # Preprocess message for chatbot output
+                if display_recv:
+                    new_msg = f"**{agent_name} (-> {recv})**: {new_msg}"  # Add role to the message
+                else:
+                    new_msg = f"**{agent_name}**: {new_msg}"
+
+                # Convert to Gradio Chatbot format: list of [user_message, bot_message] tuples
+                if agent_name == "Moderator":
+                    chatbot_output.append([new_msg, None])
+                else:
+                    chatbot_output.append([None, new_msg])
+        except Exception as e:
+            print(f"Error in _convert_to_chatbot_output: {e}")
+            print(f"all_messages type: {type(all_messages)}")
+            print(f"all_messages content: {all_messages}")
+            # Return empty list as fallback
+            return []
+        return chatbot_output
+    
+    def _create_arena_config_from_components(all_comps: dict):
+        
+        env_desc = const.GLOBAL_PROMPT
+        paper_pdf_path = all_comps[upload_file_box]
+        
+        # Step 1: Initialize the players
+        num_players = MAX_NUM_PLAYERS
+        
+        # You can ignore these fields for the demo
+        conference = "EMNLP2024"
+        paper_decision = "Accept"
+        data_dir = ''
+        paper_id = "12345"
+
+        args = Namespace(openai_client_type="openai",
+                         experiment_name="test",
+                         max_num_words=16384)
+        
+        # Phase I: Only reviewers are active
+        players = []
+        
+        # 不能直接获取role_desc，需要根据Intention_config, Knowledge_config, Responsibility_config生成一个配置
+        # self.environment.experiment_setting["players"]['Reviewer'][reviewer_index - 1]
+        
+        experiment_setting = {
+        "paper_id": paper_id,
+        "paper_decision": paper_decision,
+        "players": {
+
+            # Paper Extractor is a special player that extracts a paper from the dataset.
+            # Its constructor does not take any arguments.
+            "Paper Extractor": [{}],
+
+            # Reviewer settings are generated based on reviewer types provided in the settings.
+            "Reviewer": [],
+        },
+            # "global_settings": setting['global_settings']
+        }
+        
+        
+        for i in range(num_players):
+
+            role_name = role_desc = temperature = max_tokens = None
+
+            # All players are reviewers in Phase I only
+            role_name, expertise_config, paper_type, temperature, max_tokens = (
+                all_comps[c]
+                for c in players_idx2comp[i]
+                if not isinstance(c, (gr.Accordion, gr.Tab))
+            )
+            
+            # Use default backend
+            backend_type = DEFAULT_BACKEND
+            
+            # Use default values for other attributes since we removed them from UI
+            experiment_setting["players"]['Reviewer'].append({"is_benign": None, 
+                                                                "is_knowledgeable": None, 
+                                                                "is_responsible": None,
+                                                                "knows_authors": 'unfamous'})
+                
+            role_desc = get_reviewer_description(expertise=expertise_config, paper_type=paper_type)
+
+
+
+            # common config for all players
+            player_config = {
+                "name": role_name,
+                "role_desc": role_desc,
+                "global_prompt": env_desc,
+                "backend": {
+                    "backend_type": backend_type,
+                    "temperature": temperature,
+                    "max_tokens": max_tokens,
+                },
+            }
+            
+            player_config = AgentConfig(**player_config)
+            
+            # All players are reviewers in Phase I only
+            player = Reviewer(data_dir=data_dir, conference=conference, args=args, **player_config)
+            
+            players.append(player)
+            
+        # Phase I: Create paper extractor for reviewer assessment
+            
+        # Add paper extractor
+        paper_extractor_config = get_paper_extractor_config(max_tokens=2048)
+
+        paper_extractor = PaperExtractorPlayer(paper_pdf_path=paper_pdf_path,
+                                            data_dir=data_dir, paper_id=paper_id,
+                                            paper_decision=paper_decision, args=args,
+                                            conference=conference, **paper_extractor_config)
+        players.append(paper_extractor)
+            
+        
+        player_names = [player.name for player in players]
+        
+        # Step 2: Initialize the environment
+        env = PaperReview(player_names=player_names, paper_decision=paper_decision, paper_id=paper_id,
+                          args=args, experiment_setting=experiment_setting)
+        
+        # Step 3: Initialize the Arena
+        arena = PaperReviewArena(players=players, environment=env, args=args, global_prompt=env_desc)
+        
+        return arena
+        
+    def step_game(all_comps: dict):
+        global CURRENT_STEP_INDEX
+
+        yield {
+            btn_step: gr.update(value="Running...", interactive=False),
+            btn_restart: gr.update(interactive=False),
+        }
+        
+        cur_state = all_comps[state]
+        
+        # If arena is not yet created, create it
+        if cur_state["arena"] is None:
+            # Create the Arena
+            arena = _create_arena_config_from_components(all_comps)
+            cur_state["arena"] = arena
+        else:
+            arena = cur_state["arena"]
+        
+        # TODO: 连续运行
+        
+        timestep = arena.step()
+        
+        CURRENT_STEP_INDEX = int(arena.environment.phase_index)
+        
+        # 更新前端信息
+        if timestep:
+            try:
+                all_messages = timestep.observation
+
+                # Initialize update dictionary
+                update_dict = {
+                    btn_step: gr.update(
+                        value="Next Step", interactive=not timestep.terminal
+                    ),
+                    btn_restart: gr.update(interactive=True),
+                    state: cur_state,
+                }
+            except Exception as e:
+                print(f"Error updating main chatbot: {e}")
+                # Initialize update dictionary with empty chatbot output
+                update_dict = {
+                    btn_step: gr.update(
+                        value="Next Step", interactive=not timestep.terminal
+                    ),
+                    btn_restart: gr.update(interactive=True),
+                    state: cur_state,
+                }
+
+            # Define a mapping of player names to their respective chatbots
+            player_name_to_chatbot = {
+                "Reviewer 1": player_chatbots[0],
+                "Reviewer 2": player_chatbots[1],
+                "Reviewer 3": player_chatbots[2]
+            }
+
+            # Update each player's chatbot output
+            for player in arena.players:
+                player_name = player.name
+                if player_name in player_name_to_chatbot:
+                    try:
+                        player_messages = arena.environment.get_messages_from_player(player_name)
+                        # player_messages[0].content = 'Paper content has been extracted.'
+                        player_output = _convert_to_chatbot_output(player_messages)
+                        update_dict[player_name_to_chatbot[player_name]] = player_output
+                    except Exception as e:
+                        print(f"Error updating player {player_name} chatbot: {e}")
+                        # Set empty output as fallback
+                        update_dict[player_name_to_chatbot[player_name]] = []
+
+            # # Reviewer 1, 2, 3 Area Chair, Paper Extractor, Author
+            # for i, player in enumerate(arena.players):
+            #     player_name = player.name
+            #     # Get the messages for the current player
+            #     player_messages = arena.environment.get_observation(player_name)
+            #     player_messages[0].content = 'Paper content has been extracted.'
+            #
+            #     # Convert messages to chatbot output
+            #     player_output = _convert_to_chatbot_output(player_messages)
+
+
+                """
+                # Phase I only: All players are reviewers
+                if 'Reviewer' in player.name:
+                    player_messages = arena.environment.get_observation(player.name)
+                    # 不要显示第一条长段的信息，只显示 文章内容已被抽取
+                    player_messages[0].content = 'Paper content has been extracted.'
+                    player_output = _convert_to_chatbot_output(player_messages)
+                    # Update the player's chatbot output
+                    update_dict[player_chatbots[i]] = player_output
+                """
+            # Ahren: Auto run
+            # if not timestep.terminal:
+            #     yield from step_game(all_comps)
+
+            yield update_dict
+        
+            
+    def restart_game(all_comps: dict):
+        global CURRENT_STEP_INDEX
+        CURRENT_STEP_INDEX = 0
+        
+        cur_state = all_comps[state]
+        cur_state["arena"] = None
+        yield {
+            btn_restart: gr.update(interactive=False),
+            btn_step: gr.update(interactive=False),
+            state: cur_state,
+        }
+
+        # arena_config = _create_arena_config_from_components(all_comps)
+        # arena = Arena.from_config(arena_config)
+        # log_arena(arena, database=DB)
+        # cur_state["arena"] = arena
+
+        yield {
+            btn_step: gr.update(value="Start", interactive=True),
+            btn_restart: gr.update(interactive=True),
+            upload_file_box: gr.update(value=None),
+            state: cur_state,
+        }
+    
+    # Remove Accordion and Tab from the list of components
+    all_components = [
+        comp for comp in all_components if not isinstance(comp, (gr.Accordion, gr.Tab))
+    ]
+    
+    # update component
+
+    # If any of the Textbox, Slider, Checkbox, Dropdown, RadioButtons is changed, the Step button is disabled
+    for comp in all_components:
+
+        def _disable_step_button(state):
+            if state["arena"] is not None:
+                return gr.update(interactive=False)
+            else:
+                return gr.update()
+
+        if (
+            isinstance(
+                comp, (gr.Textbox, gr.Slider, gr.Checkbox, gr.Dropdown, gr.Radio)
+            )
+            and comp is not upload_file_box
+        ):
+            comp.change(_disable_step_button, state, btn_step)
+
+    # Add paper type change event to update all reviewer descriptions
+    def update_all_reviewer_descriptions(paper_type, *reviewer_comps):
+        # This function will be called when paper type changes
+        # For now, we'll just disable the step button
+        return gr.update(interactive=False)
+    
+    # paper_type_box.change(
+    #     fn=update_all_reviewer_descriptions,
+    #     inputs=[paper_type_box] + [comp for comp in all_components if "Expertise_config" in str(comp)],
+    #     outputs=[btn_step]
+    # )
+
+    # Auto-detect paper type functionality
+    # Get all paper type dropdowns from all reviewers
+    all_paper_type_boxes = []
+    for i in range(MAX_NUM_PLAYERS):
+        if i < DEFAULT_NUM_PLAYERS:
+            # Get the paper_type_box from each reviewer's components
+            paper_type_box = players_idx2comp[i][2]  # Index 2 is paper_type_box
+            all_paper_type_boxes.append(paper_type_box)
+    
+    # Set up auto-detect button event
+    def auto_detect_and_update(pdf_file):
+        detected_type = auto_detect_paper_type(pdf_file)
+        status_text = update_paper_type_status(detected_type)
+        # Return individual values for each dropdown and the status
+        return [detected_type] * len(all_paper_type_boxes) + [status_text]
+    
+    # Auto-detect when file is uploaded
+    upload_file_box.change(
+        fn=auto_detect_and_update,
+        inputs=[upload_file_box],
+        outputs=all_paper_type_boxes + [paper_type_status]
+    )
+    
+    # Manual auto-detect button
+    auto_detect_btn.click(
+        fn=auto_detect_and_update,
+        inputs=[upload_file_box],
+        outputs=all_paper_type_boxes + [paper_type_status]
+    )
+    
+    # Ahren: Auto run
+    btn_step.click(
+        step_game,
+        set(all_components + [state]),
+        [*player_chatbots, btn_step, btn_restart, state, upload_file_box],
+    )
+    
+    btn_restart.click(
+        restart_game,
+        set(all_components + [state]),
+        [*player_chatbots, btn_step, btn_restart, state, upload_file_box],
+    )
+
+    
+demo.queue()
+demo.launch()

autoreview/__init__.py

@@ -0,0 +1,8 @@
+import os
+
+ROOT_DIR = (
+    os.path.abspath(os.path.join(os.path.dirname(__file__), os.pardir)) + os.path.sep
+)
+EXAMPLES_DIR = os.path.join(ROOT_DIR, "examples")
+
+__version__ = "0.1.16"

autoreview/agent.py

@@ -0,0 +1,255 @@
+import logging
+import re
+import uuid
+from abc import abstractmethod
+from argparse import Namespace
+from typing import List, Union
+
+from tenacity import RetryError
+
+from .backends import IntelligenceBackend, load_backend
+from .config import AgentConfig, BackendConfig, Configurable
+from .message import SYSTEM_NAME, Message
+
+# A special signal sent by the player to indicate that it is not possible to continue the conversation, and it requests to end the conversation.
+# It contains a random UUID string to avoid being exploited by any of the players.
+SIGNAL_END_OF_CONVERSATION = f"<<<<<<END_OF_CONVERSATION>>>>>>{uuid.uuid4()}"
+
+
+class Agent(Configurable):
+    """An abstract base class for all the agents in the chatArena environment."""
+
+    @abstractmethod
+    def __init__(
+        self, name: str, role_desc: str, global_prompt: str = None, *args, **kwargs
+    ):
+        """
+        Initialize the agent.
+
+        Parameters:
+            name (str): The name of the agent.
+            role_desc (str): Description of the agent's role.
+            global_prompt (str): A universal prompt that applies to all agents. Defaults to None.
+        """
+        super().__init__(
+            name=name, role_desc=role_desc, global_prompt=global_prompt, **kwargs
+        )
+        self.name = name
+        self.role_desc = role_desc
+        self.global_prompt = global_prompt
+
+
+class Player(Agent):
+    """
+    The Player class represents a player in the chatArena environment.
+
+    A player can observe the environment
+    and perform an action (generate a response) based on the observation.
+    """
+
+    def __init__(
+        self,
+        name: str,
+        role_desc: str,
+        backend: Union[BackendConfig, IntelligenceBackend],
+        global_prompt: str = None,
+        args: Namespace = None,
+        **kwargs,
+    ):
+        """
+        Initialize the player with a name, role description, backend, and a global prompt.
+
+        Parameters:
+            name (str): The name of the player.
+            role_desc (str): Description of the player's role.
+            backend (Union[BackendConfig, IntelligenceBackend]): The backend that will be used for decision making. It can be either a LLM backend or a Human backend.
+            global_prompt (str): A universal prompt that applies to all players. Defaults to None.
+        """
+
+        self.data_dir = kwargs.pop("data_dir", None)
+        self.args = args
+
+
+        if isinstance(backend, BackendConfig):
+            backend_config = backend
+            backend_config['openai_client_type'] = args.openai_client_type
+            backend = load_backend(backend_config)
+        elif isinstance(backend, IntelligenceBackend):
+            backend_config = backend.to_config()
+        else:
+            raise ValueError(
+                f"backend must be a BackendConfig or an IntelligenceBackend, but got {type(backend)}"
+            )
+
+        assert (
+            name != SYSTEM_NAME
+        ), f"Player name cannot be {SYSTEM_NAME}, which is reserved for the system."
+
+        # Register the fields in the _config
+        super().__init__(
+            name=name,
+            role_desc=role_desc,
+            backend=backend_config,
+            global_prompt=global_prompt,
+            **kwargs,
+        )
+
+        self.backend = backend
+
+    def to_config(self) -> AgentConfig:
+        return AgentConfig(
+            name=self.name,
+            role_desc=self.role_desc,
+            backend=self.backend.to_config(),
+            global_prompt=self.global_prompt,
+        )
+
+    def act(self, observation: List[Message]) -> str:
+        """
+        Take an action based on the observation (Generate a response), which can later be parsed to actual actions that affect the game dynamics.
+
+        Parameters:
+            observation (List[Message]): The messages that the player has observed from the environment.
+
+        Returns:
+            str: The action (response) of the player.
+        """
+        try:
+            response = self.backend.query(
+                agent_name=self.name,
+                role_desc=self.role_desc,
+                history_messages=observation,
+                global_prompt=self.global_prompt,
+                request_msg=None,
+            )
+        except RetryError as e:
+            err_msg = f"Agent {self.name} failed to generate a response. Error: {e.last_attempt.exception()}. Sending signal to end the conversation."
+            logging.warning(err_msg)
+            response = SIGNAL_END_OF_CONVERSATION + err_msg
+
+        return response
+
+    def __call__(self, observation: List[Message]) -> str:
+        return self.act(observation)
+
+    async def async_act(self, observation: List[Message]) -> str:
+        """
+        Async version of act().
+
+        This is used when you want to generate a response asynchronously.
+
+        Parameters:
+            observation (List[Message]): The messages that the player has observed from the environment.
+
+        Returns:
+            str: The action (response) of the player.
+        """
+        try:
+            response = self.backend.async_query(
+                agent_name=self.name,
+                role_desc=self.role_desc,
+                history_messages=observation,
+                global_prompt=self.global_prompt,
+                request_msg=None,
+            )
+        except RetryError as e:
+            err_msg = f"Agent {self.name} failed to generate a response. Error: {e.last_attempt.exception()}. Sending signal to end the conversation."
+            logging.warning(err_msg)
+            response = SIGNAL_END_OF_CONVERSATION + err_msg
+
+        return response
+
+    def reset(self):
+        """
+        Reset the player's backend in case they are not stateless.
+
+        This is usually called at the end of each episode.
+        """
+        self.backend.reset()
+
+
+class Moderator(Player):
+    """
+    The Moderator class represents a special type of player that moderates the conversation.
+
+    It is usually used as a component of the environment when the transition dynamics is conditioned on natural language that are not easy to parse programmatically.
+    """
+
+    def __init__(
+        self,
+        role_desc: str,
+        backend: Union[BackendConfig, IntelligenceBackend],
+        terminal_condition: str,
+        global_prompt: str = None,
+        **kwargs,
+    ):
+        """
+        Initialize the moderator with a role description, backend, terminal condition, and a global prompt.
+
+        Parameters:
+            role_desc (str): Description of the moderator's role.
+            backend (Union[BackendConfig, IntelligenceBackend]): The backend that will be used for decision making.
+            terminal_condition (str): The condition that signifies the end of the conversation.
+            global_prompt (str): A universal prompt that applies to the moderator. Defaults to None.
+        """
+        name = "Moderator"
+        super().__init__(
+            name=name,
+            role_desc=role_desc,
+            backend=backend,
+            global_prompt=global_prompt,
+            **kwargs,
+        )
+
+        self.terminal_condition = terminal_condition
+
+    def to_config(self) -> AgentConfig:
+        return AgentConfig(
+            name=self.name,
+            role_desc=self.role_desc,
+            backend=self.backend.to_config(),
+            terminal_condition=self.terminal_condition,
+            global_prompt=self.global_prompt,
+        )
+
+    def is_terminal(self, history: List[Message], *args, **kwargs) -> bool:
+        """
+        Check whether an episode is terminated based on the terminal condition.
+
+        Parameters:
+            history (List[Message]): The conversation history.
+
+        Returns:
+            bool: True if the conversation is over, otherwise False.
+        """
+        # If the last message is the signal, then the conversation is over
+        if history[-1].content == SIGNAL_END_OF_CONVERSATION:
+            return True
+
+        try:
+            request_msg = Message(
+                agent_name=self.name, content=self.terminal_condition, turn=-1
+            )
+            response = self.backend.query(
+                agent_name=self.name,
+                role_desc=self.role_desc,
+                history_messages=history,
+                global_prompt=self.global_prompt,
+                request_msg=request_msg,
+                *args,
+                **kwargs,
+            )
+        except RetryError as e:
+            logging.warning(
+                f"Agent {self.name} failed to generate a response. "
+                f"Error: {e.last_attempt.exception()}."
+            )
+            return True
+
+        if re.match(
+            r"yes|y|yea|yeah|yep|yup|sure|ok|okay|alright", response, re.IGNORECASE
+        ):
+            # print(f"Decision: {response}. Conversation is ended by moderator.")
+            return True
+        else:
+            return False

autoreview/arena.py

@@ -0,0 +1,201 @@
+import csv
+import json
+import logging
+import uuid
+from typing import Dict, List, Union
+
+from .agent import Player
+from .backends import Human
+from .config import ArenaConfig
+from .environments import Environment, TimeStep, load_environment
+
+
+class TooManyInvalidActions(Exception):
+    pass
+
+
+class Arena:
+    """Utility class that manages the game environment and players."""
+
+    def __init__(
+        self, players: List[Player], environment: Environment, args,  global_prompt: str = None
+    ):
+        # Create a container for the players and environment and reset the game
+        self.players = players
+        self.environment = environment
+        self.global_prompt = global_prompt
+
+        self.current_timestep = environment.reset()
+        self.uuid = uuid.uuid4()  # Generate a unique id for the game
+        self.invalid_actions_retry = 5
+        self.args = args
+
+    @property
+    def num_players(self):
+        return self.environment.num_players
+
+    @property
+    def name_to_player(self) -> Dict[str, Player]:
+        return {player.name: player for player in self.players}
+
+    def reset(self) -> TimeStep:
+        # Reset the environment
+        self.current_timestep = self.environment.reset()
+        # Reset the players
+        for player in self.players:
+            player.reset()
+        # Reset the uuid
+        self.uuid = uuid.uuid4()
+        return self.current_timestep
+
+    def step(self) -> TimeStep:
+        """Take a step in the game: one player takes an action and the environment updates."""
+        player_name = self.environment.get_next_player()
+        player = self.name_to_player[player_name]  # get the player object
+        observation = self.environment.get_observation(
+            player_name
+        )  # get the observation for the player
+
+        timestep = None
+        for i in range(
+            self.invalid_actions_retry
+        ):  # try to take an action for a few times
+            action = player(observation)  # take an action
+            if self.environment.check_action(action, player_name):  # action is valid
+                timestep = self.environment.step(
+                    player_name, action
+                )  # update the environment
+                break
+            else:  # action is invalid
+                logging.warning(f"{player_name} made an invalid action {action}")
+                continue
+
+        if (
+            timestep is None
+        ):  # if the player made invalid actions for too many times, terminate the game
+            warning_msg = f"{player_name} has made invalid actions for {self.invalid_actions_retry} times. Terminating the game."
+            logging.warning(warning_msg)
+            raise TooManyInvalidActions(warning_msg)
+
+        return timestep
+
+    def next_is_human(self):
+        """Check if the next player is human."""
+        player_name = self.environment.get_next_player()
+        player = self.name_to_player[player_name]
+        return isinstance(player.backend, Human)
+
+    def run(self, num_steps: int = 1):
+        """Run the game for num_turns."""
+        for i in range(num_steps):
+            timestep = self.step()
+            if timestep.terminal:
+                break
+
+    @classmethod
+    def from_config(cls, config: Union[str, ArenaConfig]):
+        """Create an arena from a config."""
+        # If config is a path, load the config
+        if isinstance(config, str):
+            config = ArenaConfig.load(config)
+
+        global_prompt = config.get("global_prompt", None)
+
+        # Create the players
+        players = []
+        for player_config in config.players:
+            # Add public_prompt to the player config
+            if global_prompt is not None:
+                player_config["global_prompt"] = global_prompt
+
+            player = Player.from_config(player_config)
+            players.append(player)
+
+        # Check that the player names are unique
+        player_names = [player.name for player in players]
+        assert len(player_names) == len(
+            set(player_names)
+        ), "Player names must be unique"
+
+        # Create the environment
+        config.environment[
+            "player_names"
+        ] = player_names  # add the player names to the environment config
+        env = load_environment(config.environment)
+
+        return cls(players, env, global_prompt=global_prompt)
+
+    def to_config(self) -> ArenaConfig:
+        """Convert the arena to a config."""
+        # return {
+        #     "players": [player.to_config() for player in self.players],
+        #     "environment": self.environment.to_config(),
+        #     "global_prompt": self.global_prompt
+        # }
+        return ArenaConfig(
+            players=[player.to_config() for player in self.players],
+            environment=self.environment.to_config(),
+            global_prompt=self.global_prompt,
+        )
+
+    def launch_cli(self, max_steps: int = None, interactive: bool = True):
+        """Launch the command line interface."""
+        from autoreview.ui.cli import ArenaCLI
+
+        cli = ArenaCLI(self)
+        cli.launch(max_steps=max_steps, interactive=interactive)
+
+    def save_config(self, path: str):
+        """Save the config to a file."""
+        config = self.to_config()
+        config.save(path)
+
+    def save_history(self, path: str):
+        """
+        Save the history of the game to a file.
+
+        Supports csv and json formats.
+        """
+        messages = self.environment.get_observation()
+        message_rows = []
+
+        if path.endswith(".csv"):
+            header = [
+                "agent_name",
+                "content",
+                "turn",
+                "timestamp",
+                "visible_to",
+                "msg_type",
+            ]
+            for message in messages:
+                message_row = [
+                    message.agent_name,
+                    message.content,
+                    message.turn,
+                    str(message.timestamp),
+                    message.visible_to,
+                    message.msg_type,
+                ]
+                message_rows.append(message_row)
+
+            with open(path, "w") as f:
+                writer = csv.writer(f)
+                writer.writerow(header)
+                writer.writerows(message_rows)
+        elif path.endswith(".json"):
+            for message in messages:
+                message_row = {
+                    "agent_name": message.agent_name,
+                    "content": message.content,
+                    "turn": message.turn,
+                    "timestamp": str(message.timestamp),
+                    "visible_to": message.visible_to,
+                    "msg_type": message.msg_type,
+                }
+                message_rows.append(message_row)
+
+            with open(path, "w") as f:
+                json.dump(message_rows, f, indent=2)
+        else:
+            raise ValueError("Invalid file format")

autoreview/arguments.py

@@ -0,0 +1,171 @@
+import argparse
+import logging
+import os
+
+logger = logging.getLogger(__name__)
+
+
+def parse_args():
+    parser = argparse.ArgumentParser(description="Argument parser for configuring OpenAI API and experiment settings")
+
+    # Authentication details for OpenAI API
+    parser.add_argument(
+        "--openai_key", type=str, default=None,
+        help="API key to authenticate with OpenAI. Can be set via this argument or through the OPENAI_API_KEY environment variable."
+    )
+
+    parser.add_argument(
+        "--deployment", type=str, default=None,
+        help="For Azure OpenAI: the deployment name to be used when calling the API."
+    )
+
+    parser.add_argument(
+        "--openai_client_type", type=str, default="openai", choices=["openai", "azure_openai"],
+        help="Specify the OpenAI client type to use: 'openai' for standard OpenAI API or 'azure_openai' for Azure-hosted OpenAI services."
+    )
+
+    parser.add_argument(
+        "--endpoint", type=str, default=None,
+        help="For Azure OpenAI: custom endpoint to access the API. Should be in the format 'https://<your-endpoint>.openai.azure.com'."
+    )
+
+    parser.add_argument(
+        "--api_version", type=str, default="2023-05-15", help="API version to be used for making requests. Required "
+                                                              "for Azure OpenAI clients."
+    )
+
+    # Experiment configuration
+    parser.add_argument(
+        "--ac_scoring_method", type=str, default="ranking", choices=["recommendation", "ranking"],
+        help="Specifies the scoring method used by the Area Chair (AC) to evaluate papers: 'recommendation' or 'ranking'."
+    )
+
+    parser.add_argument(
+        "--conference", type=str, default="ICLR2023",
+        help="Conference name where the papers are being evaluated, e.g., 'ICLR2023'."
+    )
+
+    parser.add_argument(
+        "--num_reviewers_per_paper", type=int, default=3, help="The number of reviewers assigned to each paper."
+    )
+
+    parser.add_argument(
+        "--experiment_name",
+        type=str, default=None, required=False,
+        help="Specifies the name of the experiment to run. Choose from predefined experiment types based on the reviewer and AC behavior or experiment configuration."
+    )
+
+    parser.add_argument(
+        "--overwrite", action="store_true",
+        help="If set, existing results or output files will be overwritten without prompting."
+    )
+    parser.add_argument(
+        "--skip_logging", action="store_true", help="If set, we do not log the messages in the console."
+    )
+
+    parser.add_argument(
+        "--num_papers_per_area_chair", type=int, default=10,
+        help="The number of papers each area chair is assigned for evaluation."
+    )
+
+    # Model configuration
+    parser.add_argument(
+        "--model_name", type=str, default="gpt-4o", choices=["gpt-4", "gpt-4o", "gpt-35-turbo"],
+        help="Specifies which GPT model to use: 'gpt-4' for the standard GPT-4 model, 'gpt-35-turbo' for a "
+             "cost-effective alternative, or 'gpt-4o' for larger context support."
+    )
+
+    # Output directories
+    parser.add_argument(
+        "--output_dir", type=str, default="outputs", help="Directory where results, logs, and outputs will be stored."
+    )
+
+    # Output directories
+    parser.add_argument(
+        "--max_num_words", type=int, default=16384, help="Maximum number of words in the paper."
+    )
+
+    parser.add_argument(
+        "--visual_dir", type=str, default="outputs/visual",
+        help="Directory where visualization files (such as graphs and plots) will be stored."
+    )
+
+    # System configuration
+    parser.add_argument(
+        "--device", type=str, default='cuda',
+        help="The device to be used for processing (e.g., 'cuda' for GPU acceleration or 'cpu' for standard processing)."
+    )
+
+    parser.add_argument(
+        "--data_dir", type=str, default='data', help="Directory where input data (e.g., papers) are stored."
+    )
+
+    parser.add_argument(
+        "--acceptance_rate", type=float, default=0.32,
+        help="Percentage of papers to accept. We use 0.32, the average acceptance rate for ICLR 2020 - 2023"
+    )
+
+    args = parser.parse_args()
+
+    # Ensure necessary directories exist
+    os.makedirs(args.visual_dir, exist_ok=True)
+    os.makedirs(args.output_dir, exist_ok=True)
+
+    # Set 'player_to_test' based on experiment name
+    if args.experiment_name is None:
+        args.player_to_test = None
+    elif "Rx" in args.experiment_name:
+        args.player_to_test = "Reviewer"
+    elif "ACx" in args.experiment_name:
+        args.player_to_test = "Area Chair"
+    elif "no_rebuttal" in args.experiment_name or "no_overall_score" in args.experiment_name:
+        args.player_to_test = "Review Mechanism"
+
+    # Sanity checks for authentication
+    print("Running sanity checks for the arguments...")
+
+    if args.openai_client_type == "openai":
+        if os.environ.get('OPENAI_API_KEY') is None:
+            assert isinstance(args.openai_key, str), ("Please specify the `--openai_key` argument OR set the "
+                                                      "OPENAI_API_KEY environment variable.")
+            raise ValueError("OpenAI key is missing.")
+
+    EXISTING_EXPERIMENT_SETTINGS = [
+        "BASELINE", "benign_Rx1", "malicious_Rx1", "malicious_Rx2", "malicious_Rx3", "unknowledgeable_Rx1",
+        "knowledgeable_Rx1", "responsible_Rx1", "irresponsible_Rx1", "irresponsible_Rx2", "irresponsible_Rx3",
+        "inclusive_ACx1", "authoritarian_ACx1", "conformist_ACx1", "no_numeric_ratings"]
+
+    if args.experiment_name not in EXISTING_EXPERIMENT_SETTINGS:
+        logger.warning(f"Experiment name '{args.experiment_name}' is not recognized. "
+                       f"This can happen if you are customizing your own experiment settings. "
+                       f"Otherwise, please choose from the following: "
+                       f"{EXISTING_EXPERIMENT_SETTINGS}")
+
+    if args.openai_client_type == "azure_openai":
+        if os.environ.get('AZURE_OPENAI_KEY') is None:
+            assert isinstance(args.openai_key, str), ("Please specify the `--openai_key` argument OR set the "
+                                                      "AZURE_OPENAI_KEY environment variable.")
+            os.environ['AZURE_OPENAI_KEY'] = args.openai_key
+
+        if os.environ.get('AZURE_DEPLOYMENT') is None:
+            assert isinstance(args.deployment, str), ("Please specify the `--deployment` argument OR set the "
+                                                      "AZURE_DEPLOYMENT environment variable.")
+            os.environ['AZURE_DEPLOYMENT'] = args.deployment
+
+        if os.environ.get('AZURE_ENDPOINT') is None:
+            assert isinstance(args.endpoint, str), ("Please specify the `--endpoint` argument OR set the "
+                                                    "AZURE_ENDPOINT environment variable.")
+            endpoint = args.endpoint
+        else:
+            endpoint = os.environ.get('AZURE_ENDPOINT')
+
+        if not endpoint.startswith("https://"):
+            endpoint = f"https://{endpoint}.openai.azure.com"
+        os.environ['AZURE_ENDPOINT'] = endpoint
+
+        if os.environ.get('OPENAI_API_VERSION') is None:
+            assert isinstance(args.api_version, str), ("Please specify the `--api_version` argument OR set the "
+                                                       "OPENAI_API_VERSION environment variable.")
+            os.environ['OPENAI_API_VERSION'] = args.api_version
+
+    return args

autoreview/backends/__init__.py

@@ -0,0 +1,30 @@
+from ..config import BackendConfig
+from .anthropic import Claude
+from .base import IntelligenceBackend
+from .cohere import CohereAIChat
+from .hf_transformers import TransformersConversational
+from .human import Human
+from .openai import OpenAIChat
+from .dummy import Dummy
+
+ALL_BACKENDS = [
+    Human,
+    OpenAIChat,
+    CohereAIChat,
+    TransformersConversational,
+    Claude,
+    Dummy,
+]
+
+BACKEND_REGISTRY = {backend.type_name: backend for backend in ALL_BACKENDS}
+
+
+# Load a backend from a config dictionary
+def load_backend(config: BackendConfig):
+    try:
+        backend_cls = BACKEND_REGISTRY[config.backend_type]
+    except KeyError:
+        raise ValueError(f"Unknown backend type: {config.backend_type}")
+
+    backend = backend_cls.from_config(config)
+    return backend