### Clone Repository and Install LaTeXTrans

Source: https://github.com/niutrans/latextrans/blob/main/README.md

Clone the LaTeXTrans repository and install it using pip. This is the initial setup step for using the tool.

```bash
git clone https://github.com/PolarisZZM/LaTeXTrans.git
cd LaTeXTrans
pip install -e .
```

--------------------------------

### Run Streamlit Directly

Source: https://github.com/niutrans/latextrans/blob/main/README.md

Starts the Streamlit application directly for the LaTeXTrans GUI. Ensure Streamlit is installed.

```bash
streamlit run src/gui/streamlit_app.py
```

--------------------------------

### Full Translation Pipeline Programmatic Example

Source: https://context7.com/niutrans/latextrans/llms.txt

Demonstrates the complete translation pipeline using the Python API. This includes downloading sources, configuring the LLM, and running the translation workflow.

```python
import os
import toml
from src.formats.latex.utils import batch_download_arxiv_tex, extract_compressed_files
from src.agents.coordinator_agent import CoordinatorAgent

def translate_papers(arxiv_ids, output_dir="./outputs"):
    # Load and configure
    config = toml.load("config/default.toml")
    config["llm_config"]["model"] = "deepseek-chat"
    config["llm_config"]["api_key"] = os.environ.get("DEEPSEEK_API_KEY")
    config["llm_config"]["base_url"] = "https://api.deepseek.com/v1/chat/completions"
    config["target_language"] = "ch"
    config["source_language"] = "en"

    # Download sources
    projects_dir = "./tex_sources"
    os.makedirs(projects_dir, exist_ok=True)
    os.makedirs(output_dir, exist_ok=True)

    source_dirs = batch_download_arxiv_tex(arxiv_ids, projects_dir)
    extract_compressed_files(projects_dir)

    # Translate each project
    results = []
    for project_dir in source_dirs:
        try:
            coordinator = CoordinatorAgent(
                config=config,
                project_dir=project_dir,
                output_dir=output_dir
            )
            coordinator.workflow_latextrans()
            results.append({"project": project_dir, "status": "success"})
        except Exception as e:
            results.append({"project": project_dir, "status": "failed", "error": str(e)})

    return results

# Usage
results = translate_papers(["2508.18791", "2407.01648"])
for r in results:
    print(f"{r['project']}: {r['status']}")
```

--------------------------------

### LLM Provider Configuration

Source: https://context7.com/niutrans/latextrans/llms.txt

Configuration examples for different LLM providers, including DeepSeek, OpenAI, Google Gemini, and Qwen. Ensure the correct base_url is set for your chosen provider.

```toml
# DeepSeek (Recommended - cost-effective)
[llm_config]
model = "deepseek-chat"
base_url = "https://api.deepseek.com/v1/chat/completions"

# OpenAI GPT-4o
[llm_config]
model = "gpt-4o"
base_url = "https://api.openai.com/v1/chat/completions"

# Google Gemini
[llm_config]
model = "gemini-2.5-pro"
base_url = "https://generativelanguage.googleapis.com/v1beta/openai/chat/completions"

# Qwen (via compatible API)
[llm_config]
model = "qwen-3-14b"
base_url = "https://your-qwen-endpoint/v1/chat/completions"
```

--------------------------------

### Install LaTeXTrans with Conda Environment

Source: https://github.com/niutrans/latextrans/blob/main/README.md

Set up a Conda environment for LaTeXTrans to manage dependencies. This includes creating and activating the environment before cloning and installing.

```bash
conda create -n latextrans python=3.10 -y
conda activate latextrans
git clone https://github.com/PolarisZZM/LaTeXTrans.git
cd LaTeXTrans
pip install -e .
```

--------------------------------

### Install LaTeXTrans via Pip

Source: https://context7.com/niutrans/latextrans/llms.txt

Clone the repository and install LaTeXTrans in editable mode for development. An optional step to create and activate a dedicated conda environment is also provided.

```bash
# Clone the repository
git clone https://github.com/PolarisZZM/LaTeXTrans.git
cd LaTeXTrans

# Install in editable mode
pip install -e .

# Optional: Create a dedicated conda environment
conda create -n latextrans python=3.10 -y
conda activate latextrans
pip install -e .
```

--------------------------------

### Custom Terminology Configuration

Source: https://context7.com/niutrans/latextrans/llms.txt

Examples of configuring custom terminology for domain-specific translations. This can be done using a CSV file or by leveraging built-in category-based terms.

```python
# Using a CSV terminology file
# terms/custom.csv format:
# English Term,Chinese Translation
# transformer,变换器
# attention mechanism,注意力机制
# fine-tuning,微调

# Configure via config
config["user_term"] = "terms/custom.csv"

# Or use built-in category-based terms
# Terms are auto-loaded from terms/{category}.csv based on arXiv category
# e.g., terms/cs.CL.csv for computational linguistics papers
```

--------------------------------

### Get arXiv Categories

Source: https://context7.com/niutrans/latextrans/llms.txt

Retrieves arXiv categories for a given list of arXiv IDs.

```python
categories = get_arxiv_category(["2508.18791"])
# Returns: {'2508.18791': ['cs.CL', 'cs.AI']}
```

--------------------------------

### CLI: Full Command Reference and Output Structure

Source: https://context7.com/niutrans/latextrans/llms.txt

Complete CLI options for customizing translation behavior, including paths to configuration files, arXiv IDs, project paths, LLM settings, and directories. Also shows the expected output structure.

```bash
latextrans \
    --config config/default.toml \
    --arxiv 2508.18791 \
    --project /path/to/project \
    --model deepseek-chat \
    --url https://api.deepseek.com/v1/chat/completions \
    --key your-api-key \
    --source ./tex_source \
    --output ./outputs \
    --all-existing

# Expected output structure:
# outputs/
#   ch_arXiv-2508.18791v2/
#     arXiv-2508.18791v2/          # Translated LaTeX project
#       main.tex
#       ...
#     ch_arXiv-2508.18791v2.pdf    # Compiled translated PDF
#     sections_map.json            # Section mapping
#     captions_map.json            # Caption mapping
#     envs_map.json                # Environment mapping
```

--------------------------------

### Launch LaTeXTrans GUI

Source: https://github.com/niutrans/latextrans/blob/main/README.md

Launches the browser-based GUI for LaTeXTrans, offering live workflow progress, logs, and runtime configuration.

```bash
latextrans-gui
```

--------------------------------

### Launch GUI: Streamlit Interface

Source: https://context7.com/niutrans/latextrans/llms.txt

Launch the browser-based GUI for interactive translation with live progress monitoring. This can be done via the CLI entry point or by running Streamlit directly.

```bash
# Launch via CLI entry point
latextrans-gui

# Or run Streamlit directly
streamlit run src/gui/streamlit_app.py
```

--------------------------------

### Execute Latex Translation Workflow

Source: https://context7.com/niutrans/latextrans/llms.txt

Initializes and runs the complete LaTeX translation workflow using the CoordinatorAgent. Ensure configuration and directories are correctly set.

```python
config["source_language"] = "en"
config["target_language"] = "ch"

coordinator = CoordinatorAgent(
    config=config,
    project_dir="/path/to/latex/project",  # Source LaTeX project
    output_dir="./outputs"                  # Output directory
)

coordinator.workflow_latextrans()
```

--------------------------------

### Translate via Local Compressed Project

Source: https://github.com/niutrans/latextrans/blob/main/README.md

Translate a local arXiv paper project by providing the path to its compressed source package (e.g., .tar.gz).

```bash
latextrans --project D:\path\to\paper_source.tar.gz
```

--------------------------------

### CoordinatorAgent Workflow

Source: https://context7.com/niutrans/latextrans/llms.txt

Demonstrates the high-level workflow for translating LaTeX projects using the CoordinatorAgent.

```APIDOC
## CoordinatorAgent Workflow

### Description
This section outlines the steps to configure and execute the translation workflow for a LaTeX project using the `CoordinatorAgent`.

### Method
`coordinator.workflow_latextrans()`

### Endpoint
N/A (Python API)

### Parameters
#### Path Parameters
- **project_dir** (string) - Required - Source LaTeX project directory.
- **output_dir** (string) - Required - Output directory for translated files.

#### Query Parameters
None

#### Request Body
None

### Request Example
```python
config = {}
config["source_language"] = "en"
config["target_language"] = "ch"

coordinator = CoordinatorAgent(
    config=config,
    project_dir="/path/to/latex/project",
    output_dir="./outputs"
)

coordinator.workflow_latextrans()
```

### Response
#### Success Response (200)
Translated PDF at `outputs/ch_project_name/ch_project_name.pdf`.

#### Response Example
```
# Output: Translated PDF at outputs/ch_project_name/ch_project_name.pdf
```
```

--------------------------------

### Translate All Existing Projects

Source: https://github.com/niutrans/latextrans/blob/main/README.md

Process all arXiv projects currently present in the 'tex source' directory. This command is used to translate all available local projects.

```bash
latextrans --all-existing
```

--------------------------------

### Find Main TeX File

Source: https://context7.com/niutrans/latextrans/llms.txt

Finds the main TeX file in a given project directory.

```python
main_file = find_main_tex_file("/path/to/project")
# Returns: '/path/to/project/main.tex'
```

--------------------------------

### Python API: Initialize CoordinatorAgent

Source: https://context7.com/niutrans/latextrans/llms.txt

Initialize the CoordinatorAgent, the main orchestrator class for the multi-agent translation workflow. Load configuration from a TOML file and set LLM settings.

```python
import toml
from src.agents.coordinator_agent import CoordinatorAgent

# Load configuration
config = toml.load("config/default.toml")

# Configure LLM settings
config["llm_config"]["model"] = "deepseek-chat"
config["llm_config"]["api_key"] = "your-api-key"
config["llm_config"]["base_url"] = "https://api.deepseek.com/v1/chat/completions"
```

--------------------------------

### Configure LLM API Settings

Source: https://context7.com/niutrans/latextrans/llms.txt

Configure LLM API settings in `config/default.toml` before use. This includes specifying the model, API key, and base URL.

```toml
# config/default.toml
sys_name = "LaTeXTrans"
version = "0.1.0"
target_language = "ch"
source_language = "en"
paper_list = []
tex_sources_dir = "tex source"
output_dir = "outputs"
category = {}
update_term = "False"
mode = 0
user_term = ""

[llm_config]
model = "deepseek-chat"  # Model name (e.g., deepseek-chat, gpt-4o, gemini-2.5-pro)
api_key = "your_api_key_here"
base_url = "https://api.deepseek.com/v1/chat/completions"  # API endpoint
```

--------------------------------

### Generate Translated PDF with GeneratorAgent

Source: https://context7.com/niutrans/latextrans/llms.txt

Initializes and executes the GeneratorAgent to reconstruct the translated LaTeX document and compile it into a PDF. The execute method returns the path to the generated PDF or indicates failure.

```python
import toml
from src.agents.tool_agents.generator_agent import GeneratorAgent

config = toml.load("config/default.toml")

generator = GeneratorAgent(
    config=config,
    project_dir="/path/to/original/project",
    output_dir="./output"
)

pdf_path = generator.execute()

if pdf_path:
    print(f"PDF generated at: {pdf_path}")
else:
    print("PDF compilation failed")
```

--------------------------------

### CLI: Translate Local LaTeX Projects

Source: https://context7.com/niutrans/latextrans/llms.txt

Translate LaTeX projects from local directories or compressed archives. Supports processing multiple projects or all existing projects in a specified directory.

```bash
# Translate from a compressed archive
latextrans --project /path/to/paper_source.tar.gz

# Translate from an extracted directory
latextrans --project /path/to/paper_project_dir

# Process multiple local projects
latextrans --project /path/to/paper1.zip, /path/to/paper2/

# Process all existing projects in the tex source directory
latextrans --all-existing
```

--------------------------------

### Download arXiv LaTeX Sources

Source: https://context7.com/niutrans/latextrans/llms.txt

Uses the batch_download_arxiv_tex utility to download LaTeX source files from a list of arXiv IDs. Specify the save directory for the downloaded sources.

```python
from src.formats.latex.utils import (
    batch_download_arxiv_tex,
    extract_arxiv_ids,
    find_main_tex_file,
    merge_tex_from_inputs,
    remove_comments,
    extract_title,
    extract_abstract,
    get_arxiv_category
)

arxiv_ids = ["2508.18791", "2407.01648"]
source_dirs = batch_download_arxiv_tex(arxiv_ids, save_dir="./tex_sources")
```

--------------------------------

### Translate LaTeX with TranslatorAgent

Source: https://context7.com/niutrans/latextrans/llms.txt

Initializes and executes the TranslatorAgent for LLM-based translation, ensuring terminology consistency. Configure LLM model, API key, and base URL in the TOML config. Translation modes control the translation behavior.

```python
import asyncio
import toml
from src.agents.tool_agents.translator_agent import TranslatorAgent

config = toml.load("config/default.toml")
config["llm_config"]["model"] = "deepseek-chat"
config["llm_config"]["api_key"] = "your-api-key"
config["llm_config"]["base_url"] = "https://api.deepseek.com/v1/chat/completions"

translator = TranslatorAgent(
    config=config,
    project_dir="/path/to/project",
    output_dir="./output",
    trans_mode=0  # 0: Normal, 1: Retry errors, 2: With terminology
)

async def run_translation():
    await translator.execute()

asyncio.run(run_translation())
```

--------------------------------

### Translate via Local Extracted Project Directory

Source: https://github.com/niutrans/latextrans/blob/main/README.md

Translate a local arXiv paper project by providing the path to its extracted directory. This allows direct processing of unpacked source files.

```bash
latextrans --project D:\path\to\paper_project_dir
```

--------------------------------

### Configure Language Model API Keys

Source: https://github.com/niutrans/latextrans/blob/main/README.md

Edit the default TOML configuration file to set your language model's API key and base URL. Ensure these are correctly entered for the tool to function.

```toml
model = " " # model name (For example, deepseek-chat)
api_key = " " # your_api_key_here
base_url = " " # base url of the API (For example, https://api.deepseek.com/v1/chat/completions)
```

--------------------------------

### Merge TeX Inputs

Source: https://context7.com/niutrans/latextrans/llms.txt

Merges all \input and \include files from a main TeX file.

```python
full_tex = merge_tex_from_inputs("/path/to/project/main.tex")
```

--------------------------------

### CLI: Translate Single arXiv Paper

Source: https://context7.com/niutrans/latextrans/llms.txt

Translate a single arXiv paper by providing its ID. The system automatically downloads the LaTeX source, translates it, and compiles the PDF. You can also specify a version or override configuration via CLI arguments.

```bash
# Translate a single paper
latextrans --arxiv 2508.18791

# Translate a specific version
latextrans --arxiv 2508.18791v2

# Override configuration via CLI
latextrans --arxiv 2508.18791 \
    --model gpt-4o \
    --url https://api.openai.com/v1/chat/completions \
    --key sk-your-api-key \
    --output ./my_translations
```

--------------------------------

### Python API: GeneratorAgent

Source: https://context7.com/niutrans/latextrans/llms.txt

Reconstructs translated LaTeX and compiles it to PDF.

```APIDOC
## Python API: GeneratorAgent

### Description
Reconstructs translated LaTeX and compiles it to PDF.

### Method
`generator.execute()`

### Endpoint
N/A (Python API)

### Parameters
#### Path Parameters
- **project_dir** (string) - Required - Path to the original LaTeX project.
- **output_dir** (string) - Required - Output directory for the generated PDF.

#### Query Parameters
None

#### Request Body
None

### Request Example
```python
import toml
from src.agents.tool_agents.generator_agent import GeneratorAgent

config = toml.load("config/default.toml")

generator = GeneratorAgent(
    config=config,
    project_dir="/path/to/original/project",
    output_dir="./output"
)

pdf_path = generator.execute()

if pdf_path:
    print(f"PDF generated at: {pdf_path}")
else:
    print("PDF compilation failed")
```

### Response
#### Success Response (200)
Returns the path to the generated PDF file.

#### Response Example
```
PDF generated at: ./output/translated_document.pdf
```
```

--------------------------------

### Python API: Utility Functions

Source: https://context7.com/niutrans/latextrans/llms.txt

Helper functions for LaTeX processing, arXiv downloads, and file operations.

```APIDOC
## Python API: Utility Functions

### Description
Helper functions for LaTeX processing, arXiv downloads, and file operations.

### Methods
- `batch_download_arxiv_tex(arxiv_ids, save_dir)`
- `extract_arxiv_ids(mixed_input)`
- `find_main_tex_file(directory)`
- `merge_tex_from_inputs(main_tex_path, output_path)`
- `remove_comments(tex_content)`
- `extract_title(tex_content)`
- `extract_abstract(tex_content)`
- `get_arxiv_category(arxiv_id)`

### Endpoint
N/A (Python API)

### Parameters
#### Path Parameters
None

#### Query Parameters
None

#### Request Body
None

### Request Example
```python
from src.formats.latex.utils import (
    batch_download_arxiv_tex,
    extract_arxiv_ids,
    find_main_tex_file,
    merge_tex_from_inputs,
    remove_comments,
    extract_title,
    extract_abstract,
    get_arxiv_category
)

# Download arXiv sources
arxiv_ids = ["2508.18791", "2407.01648"]
source_dirs = batch_download_arxiv_tex(arxiv_ids, save_dir="./tex_sources")
# Returns: ['./tex_sources/2508.18791', './tex_sources/2407.01648']

# Extract valid arXiv IDs from mixed input
mixed_input = ["2508.18791", "https://arxiv.org/abs/2407.01648", "invalid"]
valid_ids = extract_arxiv_ids(mixed_input)
# Returns: ['2508.18791', '2407.01648']
```

### Response
#### Success Response (200)
Depends on the function called. Returns file paths, lists of IDs, or extracted content.

#### Response Example
```
# Example for batch_download_arxiv_tex:
# ['./tex_sources/2508.18791', './tex_sources/2407.01648']

# Example for extract_arxiv_ids:
# ['2508.18791', '2407.01648']
```
```

--------------------------------

### Translate via ArXiv ID

Source: https://github.com/niutrans/latextrans/blob/main/README.md

Initiate translation for a single arXiv paper by providing its ID. The tool will download, process, and save the translated output.

```bash
latextrans --arxiv ${xxxx}
# For example, 
# latextrans --arxiv 2508.18791
```

--------------------------------

### CLI: Batch Translate Multiple Papers

Source: https://context7.com/niutrans/latextrans/llms.txt

Translate multiple arXiv papers in a single run using comma-separated IDs. This can include a mix of arXiv IDs and URLs.

```bash
# Batch translate multiple papers
latextrans --arxiv 2508.18791v2, 2407.01648, 2501.12948

# Mix arXiv IDs and URLs
latextrans --arxiv 2508.18791, https://arxiv.org/abs/2407.01648
```

--------------------------------

### Parse LaTeX Documents with LatexParser

Source: https://context7.com/niutrans/latextrans/llms.txt

Initializes and uses the LatexParser to convert LaTeX files into structured JSON. Access parsed components like sections, captions, and environments from the parser's attributes.

```python
from src.formats.latex.parser import LatexParser

parser = LatexParser(
    dir="/path/to/latex/project",
    output_dir="./parsed_output"
)

parser.parse()

sections = parser.sections_json      # List of section dictionaries
captions = parser.captions_json      # List of caption dictionaries
envs = parser.envs_json              # List of environment dictionaries
newcommands = parser.newcommands_json  # List of newcommand dictionaries
inputs = parser.inputs_json          # List of input/include mappings
```

--------------------------------

### Batch Translate via ArXiv IDs

Source: https://github.com/niutrans/latextrans/blob/main/README.md

Translate multiple arXiv papers simultaneously by providing a comma-separated list of their IDs. This is useful for processing several papers at once.

```bash
latextrans --arxiv ${xxxx}, ${xxxx}
# For example,
# latextrans --arxiv 2508.18791v2, 2407.01648
```

--------------------------------

### Extract Metadata from LaTeX

Source: https://context7.com/niutrans/latextrans/llms.txt

Extracts the title and abstract from cleaned LaTeX code.

```python
title = extract_title(clean_tex)
abstract = extract_abstract(clean_tex)
```

--------------------------------

### Python API: TranslatorAgent

Source: https://context7.com/niutrans/latextrans/llms.txt

Handles LLM-based translation with terminology consistency.

```APIDOC
## Python API: TranslatorAgent

### Description
The translation agent that handles LLM-based translation with terminology consistency.

### Method
`translator.execute()`

### Endpoint
N/A (Python API)

### Parameters
#### Path Parameters
- **project_dir** (string) - Required - Path to the project directory.
- **output_dir** (string) - Required - Output directory for translated files.

#### Query Parameters
None

#### Request Body
None

### Request Example
```python
import asyncio
import toml
from src.agents.tool_agents.translator_agent import TranslatorAgent

config = toml.load("config/default.toml")
config["llm_config"]["model"] = "deepseek-chat"
config["llm_config"]["api_key"] = "your-api-key"
config["llm_config"]["base_url"] = "https://api.deepseek.com/v1/chat/completions"

translator = TranslatorAgent(
    config=config,
    project_dir="/path/to/project",
    output_dir="./output",
    trans_mode=0  # 0: Normal, 1: Retry errors, 2: With terminology
)

async def run_translation():
    await translator.execute()

asyncio.run(run_translation())
```

### Response
#### Success Response (200)
Translation process is executed.

#### Response Example
```
# Translation executed successfully.
```

### Translation Modes
- **mode=0**: Standard translation.
- **mode=1**: Retranslate error parts only.
- **mode=2**: Translation with terminology dictionary.
```

--------------------------------

### Python API: LatexParser

Source: https://context7.com/niutrans/latextrans/llms.txt

API for parsing LaTeX documents into structured JSON mappings.

```APIDOC
## Python API: LatexParser

### Description
Parses LaTeX documents into structured JSON mappings for sections, environments, captions, and commands.

### Method
`parser.parse()`

### Endpoint
N/A (Python API)

### Parameters
#### Path Parameters
- **dir** (string) - Required - Path to the LaTeX project directory.
- **output_dir** (string) - Required - Directory to save parsed output.

#### Query Parameters
None

#### Request Body
None

### Request Example
```python
from src.formats.latex.parser import LatexParser

parser = LatexParser(
    dir="/path/to/latex/project",
    output_dir="./parsed_output"
)

parser.parse()
```

### Response
#### Success Response (200)
Access parsed components via attributes:
- **sections_json**: List of section dictionaries.
- **captions_json**: List of caption dictionaries.
- **envs_json**: List of environment dictionaries.
- **newcommands_json**: List of newcommand dictionaries.
- **inputs_json**: List of input/include mappings.

#### Response Example
```json
{
    "section": "1",
    "content": "\\section{...}",
    "trans_content": ""
}

{
    "placeholder": "<PLACEHOLDER_ENV_1>",
    "env_name": "theorem",
    "content": "\\begin{theorem}...\\end{theorem}",
    "trans_content": "",
    "need_trans": true
}
```
```

--------------------------------

### Validate LaTeX Translations with ValidatorAgent

Source: https://context7.com/niutrans/latextrans/llms.txt

Initializes and executes the ValidatorAgent to check for LaTeX command integrity, placeholder preservation, and bracket matching. The execute method returns a report of errors, which can be retried.

```python
import toml
from src.agents.tool_agents.validator_agent import ValidatorAgent
from pathlib import Path

config = toml.load("config/default.toml")

validator = ValidatorAgent(
    config=config,
    project_dir="/path/to/project",
    output_dir="./output"
)

errors_report = validator.execute()

updated_errors = validator.execute(errors_report=errors_report)
```

--------------------------------

### LaTeXTrans Citation

Source: https://github.com/niutrans/latextrans/blob/main/README.md

BibTeX entry for citing the LaTeXTrans project in academic work. Includes title, authors, journal, and year.

```bibtex
@article{zhu2025latextrans,
  title={LaTeXTrans: Structured LaTeX Translation with Multi-Agent Coordination},
  author={Zhu, Ziming and Wang, Chenglong and Xing, Shunjie and Huo, Yifu and Tian, Fengning and Du, Quan and Yang, Di and Zhang, Chunliang and Xiao, Tong and Zhu, Jingbo},
  journal={arXiv preprint arXiv:2508.18791},
  year={2025}
}
```

--------------------------------

### Python API: ValidatorAgent

Source: https://context7.com/niutrans/latextrans/llms.txt

Validates translations for LaTeX command integrity, placeholder preservation, and bracket matching.

```APIDOC
## Python API: ValidatorAgent

### Description
Validates translations for LaTeX command integrity, placeholder preservation, and bracket matching.

### Method
`validator.execute()`

### Endpoint
N/A (Python API)

### Parameters
#### Path Parameters
- **project_dir** (string) - Required - Path to the project directory.
- **output_dir** (string) - Required - Output directory for validation results.

#### Query Parameters
None

#### Request Body
None

### Request Example
```python
import toml
from src.agents.tool_agents.validator_agent import ValidatorAgent
from pathlib import Path

config = toml.load("config/default.toml")

validator = ValidatorAgent(
    config=config,
    project_dir="/path/to/project",
    output_dir="./output"
)

errors_report = validator.execute()

# Retry validation with existing error report
updated_errors = validator.execute(errors_report=errors_report)
```

### Response
#### Success Response (200)
Returns a list of validation errors.

#### Response Example
```json
[
    {
        "part": "sec",
        "num_or_ph": "1",
        "command_error": "...",
        "ph_error": "...",
        "bracket_error": "..."
    }
]
```
```

--------------------------------

### Clean LaTeX Code

Source: https://context7.com/niutrans/latextrans/llms.txt

Removes comments from LaTeX code.

```python
clean_tex = remove_comments(full_tex)
```

--------------------------------

### Extract Valid arXiv IDs

Source: https://context7.com/niutrans/latextrans/llms.txt

Employs the extract_arxiv_ids utility to filter a list of inputs, returning only valid arXiv identifiers. This function handles various input formats, including direct IDs and URLs.

```python
from src.formats.latex.utils import (
    batch_download_arxiv_tex,
    extract_arxiv_ids,
    find_main_tex_file,
    merge_tex_from_inputs,
    remove_comments,
    extract_title,
    extract_abstract,
    get_arxiv_category
)

mixed_input = ["2508.18791", "https://arxiv.org/abs/2407.01648", "invalid"]
valid_ids = extract_arxiv_ids(mixed_input)
```

--------------------------------

### Translate Specific ArXiv ID Revision

Source: https://github.com/niutrans/latextrans/blob/main/README.md

Translate a specific version of an arXiv paper by appending the revision number to the arXiv ID. This allows targeting particular paper states.

```bash
latextrans --arxiv 2508.18791v2
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.