# 🛠️ robust-json
[](https://www.python.org/downloads/)
[](https://opensource.org/licenses/MIT)
**Robust JSON extraction and repair utilities for LLM-generated content.**
Parse JSON from messy LLM outputs with confidence. `robust-json` extracts and repairs JSON even when models mix commentary with structured data, use incorrect quotes, add trailing commas, include comments, or truncate responses mid-object.
---
## ✨ Why robust-json?
Large Language Models are powerful but inconsistent when generating JSON. They might:
- 📝 **Mix text and JSON**: Embed JSON inside markdown code blocks or conversational responses
- 💬 **Add comments**: Include `//` or `#` comments that break standard JSON parsers
- 🔤 **Use wrong quotes**: Generate single quotes (`'`) instead of double quotes (`"`)
- 🔚 **Add trailing commas**: Place commas after the last item in arrays/objects
- ✂️ **Truncate output**: Stop mid-JSON due to token limits or errors
`robust-json` handles all these cases automatically, so you can focus on using the data instead of fighting with parser errors.
---
## 🚀 Features
- **🔍 Smart extraction**: Automatically finds JSON objects and arrays within free-form text
- **🔧 Auto-repair**: Fixes common LLM errors including:
- Single-quoted strings → double quotes
- Mixed quote types (e.g., `'text"` → `'text'`)
- Inline comments (`//` and `#`)
- Trailing commas
- Unclosed braces and brackets
- **🎯 Multiple parsers**: Falls back through `json` → `ast.literal_eval` for maximum compatibility
- **⚡ Performance**: Optional speedups with `regex` (enhanced regex engine) and `numba` (JIT-compiled bracket scanning)
- **🌍 Unicode support**: Handles international characters and emoji seamlessly
---
## 📦 Installation
**Basic installation:**
```bash
pip install robust-json-parser
```
**With performance optimizations (numba JIT):**
```bash
pip install robust-json-parser[speedups]
```
**With regex (enhanced regex engine with better Unicode support):**
```bash
pip install robust-json-parser[regex]
```
**All extras:**
```bash
pip install robust-json-parser[speedups,regex]
```
**Requirements:** Python 3.9+
---
## 🎯 Quick Start
### Basic Usage
```python
from robust_json import loads
# LLM output with mixed formatting
llm_response = """
Sure! Here's the data you requested:
```json
{
"name": "Alice",
"age": 30,
"hobbies": ["reading", "coding",], // trailing comma
"active": true, # Python-style comment
}
Hope this helps!
"""
data = loads(llm_response)
print(data)
# {'name': 'Alice', 'age': 30, 'hobbies': ['reading', 'coding'], 'active': True}
```
### Handling Malformed JSON
```python
from robust_json import loads
# Mixed quotes, comments, and multilingual text
message = """
Hello, I'm a recruitment consultant. Here's the job description for your matching assessment:
```json
{"id": "algo", "position": "Large Language Model Algorithm Engineer",
# this is the keywords list used to analyze the candidate
"keywords": {"positive": ["PEFT", "RLHF"], "negative": ["CNN", "RNN"]}, # negative keywords is supported
"summary": 'The candidate has some AI background, but lacks experience."
}
"""
data = loads(message)
print(data["keywords"]["positive"])
# ['PEFT', 'RLHF']
```
### Truncated/Partial JSON
```python
from robust_json import loads
# JSON cut off mid-object
incomplete = '{"user": {"name": "Bob", "email": "bob@example.com"'
data = loads(incomplete)
print(data)
# {'user': {'name': 'Bob', 'email': 'bob@example.com'}}
```
### Extract Multiple JSON Objects
```python
from robust_json import extract_all, RobustJSONParser
text = """
First result: {"a": 1, "b": 2}
Some text in between...
Second result: {"x": 10, "y": 20}
"""
# Get all extractions with metadata
extractions = extract_all(text)
for extraction in extractions:
print(f"Found at position {extraction.start}: {extraction.text}")
# Or just get the parsed objects
parser = RobustJSONParser()
objects = parser.parse_all(text)
print(objects)
# [{'a': 1, 'b': 2}, {'x': 10, 'y': 20}]
```
---
## 📚 API Reference
### `loads(source, *, allow_partial=True, default=None, strict=False)`
Parse the first JSON object found in the source text.
**Parameters:**
- `source` (str): Text containing JSON
- `allow_partial` (bool): If `True`, auto-complete truncated JSON (default: `True`)
- `default` (Optional): Return this value if no JSON found (default: `None` raises error)
- `strict` (bool): If `True`, only extract from code blocks and brace-delimited content (default: `False`)
**Returns:** Parsed Python object (dict, list, etc.)
**Raises:** `ValueError` if no JSON found and no default provided
---
### `extract(source, *, allow_partial=True)`
Extract the first JSON-like fragment with metadata.
**Returns:** `Extraction` object or `None`
---
### `extract_all(source, *, allow_partial=True)`
Extract all JSON-like fragments from text.
**Returns:** List of `Extraction` objects
---
### `RobustJSONParser`
Main parser class for advanced usage.
**Methods:**
- `extract(source, limit=None)`: Find JSON fragments (returns list of `Extraction` objects)
- `parse_first(source)`: Parse first JSON object (returns parsed object or `None`)
- `parse_all(source)`: Parse all JSON objects (returns list of parsed objects)
**Parameters:**
- `allow_partial` (bool): Auto-complete truncated JSON (default: `True`)
- `strict` (bool): Only extract from explicit JSON contexts (default: `False`)
---
### `Extraction`
Dataclass representing an extracted JSON candidate.
**Attributes:**
- `text` (str): The extracted text
- `start` (int): Starting position in source
- `end` (int): Ending position in source
- `is_partial` (bool): Whether the extraction appears truncated
- `repaired` (Optional[str]): The repaired version after processing
---
## 🔧 How It Works
1. **🔎 Extraction**: Scans text for JSON patterns using:
- Markdown code blocks (`` ```json ... ``` ``)
- Brace-balanced regions (`{...}`, `[...]`)
2. **🛠️ Repair**: Applies fixes in order:
- Strip `//` and `#` comments
- Fix mixed quote types (e.g., `'text"` → `'text'`)
- Normalize single quotes to double quotes
- Remove trailing commas
- Balance unclosed braces (if `allow_partial=True`)
3. **✅ Parse**: Attempts parsing with:
- `json.loads()` (standard JSON)
- `ast.literal_eval()` (Python literals)
4. **📊 Return**: Returns first successful parse or continues to next candidate
---
## 🎨 Use Cases
- **🤖 LLM Integration**: Parse structured output from ChatGPT, Claude, Llama, etc.
- **📊 Data Extraction**: Extract JSON from logs, documentation, or mixed-format files
- **🔄 API Responses**: Handle malformed API responses gracefully
- **🧪 Testing**: Validate and repair JSON in test fixtures
- **📝 Data Migration**: Clean up inconsistent JSON during migrations
---
## ⚡ Performance Tips
1. **Install speedups** for large-scale processing:
```bash
pip install robust-json-parser[speedups] # numba JIT compilation
pip install robust-json-parser[regex] # enhanced regex engine with better Unicode support
```
2. **Use strict mode** when JSON is always in code blocks:
```python
loads(text, strict=True) # Faster, skips fallback attempts
```
3. **Disable partial completion** if you know JSON is complete:
```python
loads(text, allow_partial=False) # Skips brace-balancing step
```
4. **Reuse parser instance** for multiple parses:
```python
parser = RobustJSONParser()
for text in texts:
data = parser.parse_first(text)
```
---
## 🧪 Test Status
**Overall Test Coverage: 98.6% (140/142 tests passing)**
| Category | Test File | Passed | Failed | Total | Pass Rate | Status |
|----------|-----------|--------|--------|-------|-----------|---------|
| **Core Functionality** | test_parser.py | 5 | 0 | 5 | 100.0% | ✅ |
| **Comprehensive Tests** | test_comprehensive.py | 50 | 1 | 51 | 98.0% | ✅ |
| **Edge Cases** | test_edge_cases.py | 38 | 1 | 39 | 97.4% | ✅ |
| **LLM Scenarios** | test_llm_scenarios.py | 31 | 0 | 31 | 100.0% | ✅ |
| **Performance** | test_performance.py | 11 | 0 | 11 | 100.0% | ✅ |
| **Batch Processing** | test_batch_performance.py | 5 | 0 | 5 | 100.0% | ✅ |
### Test Categories Breakdown
- **✅ Core Functionality (100%)**: Basic parsing, extraction, and repair features
- **✅ Comprehensive Tests (98.0%)**: Real-world scenarios, complex nested structures, multilingual content
- **✅ Edge Cases (97.4%)**: Unicode handling, malformed JSON, bracket matching, error recovery
- **✅ LLM Scenarios (100%)**: ChatGPT/Claude-style outputs, conversational text extraction
- **✅ Performance (100%)**: Large datasets, memory usage, parsing speed benchmarks
- **✅ Batch Processing (100%)**: Parallel processing, multiprocessing, error handling
### Known Issues (2 failing tests)
- **Extraction Order**: `extract_all` function needs to preserve proper ordering
- **Deep Nesting**: Complex nested structures with mismatched brackets need enhanced repair
---
## 🤝 Contributing
We welcome contributions from developers of all skill levels! Whether you're fixing bugs, adding features, or improving documentation, your help makes this project better for everyone.
### 🎯 How to Contribute
1. **🐛 Bug Reports**: Found an issue? Open a GitHub issue with:
- Clear description of the problem
- Minimal reproducible example
- Expected vs actual behavior
2. **✨ Feature Requests**: Have an idea? We'd love to hear it! Open an issue to discuss:
- Use case and motivation
- Proposed implementation approach
- Any breaking changes
3. **🔧 Code Contributions**: Ready to code? Here's how:
```bash
# Fork and clone the repository
git clone https://github.com/your-username/robust-json.git
cd robust-json
# Install in development mode
pip install -e ".[speedups,regex,dev]"
# Run tests to ensure everything works
pytest tests/
# Make your changes and test them
pytest tests/ -v
# Submit a pull request
```
### 🧪 Testing Your Changes
```bash
# Run all tests
pytest tests/
# Run specific test categories
pytest tests/test_parser.py # Core functionality
pytest tests/test_comprehensive.py # Comprehensive scenarios
pytest tests/test_llm_scenarios.py # LLM-specific cases
pytest tests/test_edge_cases.py # Edge cases and error handling
pytest tests/test_performance.py # Performance benchmarks
# Run with coverage
pytest tests/ --cov=robust_json --cov-report=html
```
### 🎨 Areas We'd Love Help With
- **🌍 Internationalization**: Better support for non-Latin scripts and RTL languages
- **⚡ Performance**: Optimize parsing speed for very large JSON objects
- **🔍 LLM Integration**: Improve extraction from more LLM output formats
- **📚 Documentation**: Examples, tutorials, and API documentation
- **🧪 Test Coverage**: Add more edge cases and real-world scenarios
- **🐛 Bug Fixes**: Help us get to 100% test pass rate!
### 📋 Development Guidelines
- **Code Style**: Follow PEP 8, use type hints, and add docstrings
- **Testing**: Add tests for new features and bug fixes
- **Documentation**: Update README and docstrings as needed
- **Performance**: Consider performance impact of changes
- **Compatibility**: Maintain Python 3.9+ compatibility
### 🏆 Recognition
Contributors will be recognized in our README and release notes. We appreciate every contribution, no matter how small!
**Ready to get started?** Check out our [open issues](https://github.com/callzhang/robust-json/issues) or start with the failing tests above!
---
## 📝 License
This project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.
---
## 🙏 Acknowledgments
Built for developers working with LLM-generated content who need reliability without sacrificing flexibility.
---
**Made with ❤️ for the AI/LLM community**
Raw data
{
"_id": null,
"home_page": null,
"name": "robust-json-parser",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.9",
"maintainer_email": "Derek <derek@preseen.ai>",
"keywords": "json, llm, ai, validation, repair, parser, extraction, chatgpt, claude",
"author": null,
"author_email": "Derek <derek@preseen.ai>",
"download_url": "https://files.pythonhosted.org/packages/7b/45/a210bf643e0a8123f2c7f2c142510c34194d959e95164bbb18c18318617c/robust_json_parser-0.1.5.tar.gz",
"platform": null,
"description": "# \ud83d\udee0\ufe0f robust-json\n\n[](https://www.python.org/downloads/)\n[](https://opensource.org/licenses/MIT)\n\n**Robust JSON extraction and repair utilities for LLM-generated content.**\n\nParse JSON from messy LLM outputs with confidence. `robust-json` extracts and repairs JSON even when models mix commentary with structured data, use incorrect quotes, add trailing commas, include comments, or truncate responses mid-object.\n\n---\n\n## \u2728 Why robust-json?\n\nLarge Language Models are powerful but inconsistent when generating JSON. They might:\n\n- \ud83d\udcdd **Mix text and JSON**: Embed JSON inside markdown code blocks or conversational responses\n- \ud83d\udcac **Add comments**: Include `//` or `#` comments that break standard JSON parsers\n- \ud83d\udd24 **Use wrong quotes**: Generate single quotes (`'`) instead of double quotes (`\"`)\n- \ud83d\udd1a **Add trailing commas**: Place commas after the last item in arrays/objects\n- \u2702\ufe0f **Truncate output**: Stop mid-JSON due to token limits or errors\n\n`robust-json` handles all these cases automatically, so you can focus on using the data instead of fighting with parser errors.\n\n---\n\n## \ud83d\ude80 Features\n\n- **\ud83d\udd0d Smart extraction**: Automatically finds JSON objects and arrays within free-form text\n- **\ud83d\udd27 Auto-repair**: Fixes common LLM errors including:\n - Single-quoted strings \u2192 double quotes\n - Mixed quote types (e.g., `'text\"` \u2192 `'text'`)\n - Inline comments (`//` and `#`)\n - Trailing commas\n - Unclosed braces and brackets\n- **\ud83c\udfaf Multiple parsers**: Falls back through `json` \u2192 `ast.literal_eval` for maximum compatibility\n- **\u26a1 Performance**: Optional speedups with `regex` (enhanced regex engine) and `numba` (JIT-compiled bracket scanning)\n- **\ud83c\udf0d Unicode support**: Handles international characters and emoji seamlessly\n\n---\n\n## \ud83d\udce6 Installation\n\n**Basic installation:**\n```bash\npip install robust-json-parser\n```\n\n**With performance optimizations (numba JIT):**\n```bash\npip install robust-json-parser[speedups]\n```\n\n**With regex (enhanced regex engine with better Unicode support):**\n```bash\npip install robust-json-parser[regex]\n```\n\n**All extras:**\n```bash\npip install robust-json-parser[speedups,regex]\n```\n\n**Requirements:** Python 3.9+\n\n---\n\n## \ud83c\udfaf Quick Start\n\n### Basic Usage\n\n```python\nfrom robust_json import loads\n\n# LLM output with mixed formatting\nllm_response = \"\"\"\nSure! Here's the data you requested:\n```json\n{\n \"name\": \"Alice\",\n \"age\": 30,\n \"hobbies\": [\"reading\", \"coding\",], // trailing comma\n \"active\": true, # Python-style comment\n}\n\nHope this helps!\n\"\"\"\n\ndata = loads(llm_response)\nprint(data)\n# {'name': 'Alice', 'age': 30, 'hobbies': ['reading', 'coding'], 'active': True}\n```\n\n### Handling Malformed JSON\n\n```python\nfrom robust_json import loads\n\n# Mixed quotes, comments, and multilingual text\nmessage = \"\"\"\nHello, I'm a recruitment consultant. Here's the job description for your matching assessment:\n```json\n{\"id\": \"algo\", \"position\": \"Large Language Model Algorithm Engineer\",\n# this is the keywords list used to analyze the candidate\n \"keywords\": {\"positive\": [\"PEFT\", \"RLHF\"], \"negative\": [\"CNN\", \"RNN\"]}, # negative keywords is supported\n \"summary\": 'The candidate has some AI background, but lacks experience.\"\n }\n\"\"\"\n\ndata = loads(message)\nprint(data[\"keywords\"][\"positive\"])\n# ['PEFT', 'RLHF']\n```\n\n### Truncated/Partial JSON\n\n```python\nfrom robust_json import loads\n\n# JSON cut off mid-object\nincomplete = '{\"user\": {\"name\": \"Bob\", \"email\": \"bob@example.com\"'\n\ndata = loads(incomplete)\nprint(data)\n# {'user': {'name': 'Bob', 'email': 'bob@example.com'}}\n```\n\n### Extract Multiple JSON Objects\n\n```python\nfrom robust_json import extract_all, RobustJSONParser\n\ntext = \"\"\"\nFirst result: {\"a\": 1, \"b\": 2}\nSome text in between...\nSecond result: {\"x\": 10, \"y\": 20}\n\"\"\"\n\n# Get all extractions with metadata\nextractions = extract_all(text)\nfor extraction in extractions:\n print(f\"Found at position {extraction.start}: {extraction.text}\")\n\n# Or just get the parsed objects\nparser = RobustJSONParser()\nobjects = parser.parse_all(text)\nprint(objects)\n# [{'a': 1, 'b': 2}, {'x': 10, 'y': 20}]\n```\n\n---\n\n## \ud83d\udcda API Reference\n\n### `loads(source, *, allow_partial=True, default=None, strict=False)`\n\nParse the first JSON object found in the source text.\n\n**Parameters:**\n- `source` (str): Text containing JSON\n- `allow_partial` (bool): If `True`, auto-complete truncated JSON (default: `True`)\n- `default` (Optional): Return this value if no JSON found (default: `None` raises error)\n- `strict` (bool): If `True`, only extract from code blocks and brace-delimited content (default: `False`)\n\n**Returns:** Parsed Python object (dict, list, etc.)\n\n**Raises:** `ValueError` if no JSON found and no default provided\n\n---\n\n### `extract(source, *, allow_partial=True)`\n\nExtract the first JSON-like fragment with metadata.\n\n**Returns:** `Extraction` object or `None`\n\n---\n\n### `extract_all(source, *, allow_partial=True)`\n\nExtract all JSON-like fragments from text.\n\n**Returns:** List of `Extraction` objects\n\n---\n\n### `RobustJSONParser`\n\nMain parser class for advanced usage.\n\n**Methods:**\n- `extract(source, limit=None)`: Find JSON fragments (returns list of `Extraction` objects)\n- `parse_first(source)`: Parse first JSON object (returns parsed object or `None`)\n- `parse_all(source)`: Parse all JSON objects (returns list of parsed objects)\n\n**Parameters:**\n- `allow_partial` (bool): Auto-complete truncated JSON (default: `True`)\n- `strict` (bool): Only extract from explicit JSON contexts (default: `False`)\n\n---\n\n### `Extraction`\n\nDataclass representing an extracted JSON candidate.\n\n**Attributes:**\n- `text` (str): The extracted text\n- `start` (int): Starting position in source\n- `end` (int): Ending position in source\n- `is_partial` (bool): Whether the extraction appears truncated\n- `repaired` (Optional[str]): The repaired version after processing\n\n---\n\n## \ud83d\udd27 How It Works\n\n1. **\ud83d\udd0e Extraction**: Scans text for JSON patterns using:\n - Markdown code blocks (`` ```json ... ``` ``)\n - Brace-balanced regions (`{...}`, `[...]`)\n\n2. **\ud83d\udee0\ufe0f Repair**: Applies fixes in order:\n - Strip `//` and `#` comments\n - Fix mixed quote types (e.g., `'text\"` \u2192 `'text'`)\n - Normalize single quotes to double quotes\n - Remove trailing commas\n - Balance unclosed braces (if `allow_partial=True`)\n\n3. **\u2705 Parse**: Attempts parsing with:\n - `json.loads()` (standard JSON)\n - `ast.literal_eval()` (Python literals)\n\n4. **\ud83d\udcca Return**: Returns first successful parse or continues to next candidate\n\n---\n\n## \ud83c\udfa8 Use Cases\n\n- **\ud83e\udd16 LLM Integration**: Parse structured output from ChatGPT, Claude, Llama, etc.\n- **\ud83d\udcca Data Extraction**: Extract JSON from logs, documentation, or mixed-format files\n- **\ud83d\udd04 API Responses**: Handle malformed API responses gracefully\n- **\ud83e\uddea Testing**: Validate and repair JSON in test fixtures\n- **\ud83d\udcdd Data Migration**: Clean up inconsistent JSON during migrations\n\n---\n\n## \u26a1 Performance Tips\n\n1. **Install speedups** for large-scale processing:\n ```bash\n pip install robust-json-parser[speedups] # numba JIT compilation\n pip install robust-json-parser[regex] # enhanced regex engine with better Unicode support\n ```\n\n2. **Use strict mode** when JSON is always in code blocks:\n ```python\n loads(text, strict=True) # Faster, skips fallback attempts\n ```\n\n3. **Disable partial completion** if you know JSON is complete:\n ```python\n loads(text, allow_partial=False) # Skips brace-balancing step\n ```\n\n4. **Reuse parser instance** for multiple parses:\n ```python\n parser = RobustJSONParser()\n for text in texts:\n data = parser.parse_first(text)\n ```\n\n---\n\n## \ud83e\uddea Test Status\n\n**Overall Test Coverage: 98.6% (140/142 tests passing)**\n\n| Category | Test File | Passed | Failed | Total | Pass Rate | Status |\n|----------|-----------|--------|--------|-------|-----------|---------|\n| **Core Functionality** | test_parser.py | 5 | 0 | 5 | 100.0% | \u2705 |\n| **Comprehensive Tests** | test_comprehensive.py | 50 | 1 | 51 | 98.0% | \u2705 |\n| **Edge Cases** | test_edge_cases.py | 38 | 1 | 39 | 97.4% | \u2705 |\n| **LLM Scenarios** | test_llm_scenarios.py | 31 | 0 | 31 | 100.0% | \u2705 |\n| **Performance** | test_performance.py | 11 | 0 | 11 | 100.0% | \u2705 |\n| **Batch Processing** | test_batch_performance.py | 5 | 0 | 5 | 100.0% | \u2705 |\n\n### Test Categories Breakdown\n\n- **\u2705 Core Functionality (100%)**: Basic parsing, extraction, and repair features\n- **\u2705 Comprehensive Tests (98.0%)**: Real-world scenarios, complex nested structures, multilingual content\n- **\u2705 Edge Cases (97.4%)**: Unicode handling, malformed JSON, bracket matching, error recovery\n- **\u2705 LLM Scenarios (100%)**: ChatGPT/Claude-style outputs, conversational text extraction\n- **\u2705 Performance (100%)**: Large datasets, memory usage, parsing speed benchmarks\n- **\u2705 Batch Processing (100%)**: Parallel processing, multiprocessing, error handling\n\n### Known Issues (2 failing tests)\n- **Extraction Order**: `extract_all` function needs to preserve proper ordering\n- **Deep Nesting**: Complex nested structures with mismatched brackets need enhanced repair\n\n---\n\n## \ud83e\udd1d Contributing\n\nWe welcome contributions from developers of all skill levels! Whether you're fixing bugs, adding features, or improving documentation, your help makes this project better for everyone.\n\n### \ud83c\udfaf How to Contribute\n\n1. **\ud83d\udc1b Bug Reports**: Found an issue? Open a GitHub issue with:\n - Clear description of the problem\n - Minimal reproducible example\n - Expected vs actual behavior\n\n2. **\u2728 Feature Requests**: Have an idea? We'd love to hear it! Open an issue to discuss:\n - Use case and motivation\n - Proposed implementation approach\n - Any breaking changes\n\n3. **\ud83d\udd27 Code Contributions**: Ready to code? Here's how:\n ```bash\n # Fork and clone the repository\n git clone https://github.com/your-username/robust-json.git\n cd robust-json\n \n # Install in development mode\n pip install -e \".[speedups,regex,dev]\"\n \n # Run tests to ensure everything works\n pytest tests/\n \n # Make your changes and test them\n pytest tests/ -v\n \n # Submit a pull request\n ```\n\n### \ud83e\uddea Testing Your Changes\n\n```bash\n# Run all tests\npytest tests/\n\n# Run specific test categories\npytest tests/test_parser.py # Core functionality\npytest tests/test_comprehensive.py # Comprehensive scenarios\npytest tests/test_llm_scenarios.py # LLM-specific cases\npytest tests/test_edge_cases.py # Edge cases and error handling\npytest tests/test_performance.py # Performance benchmarks\n\n# Run with coverage\npytest tests/ --cov=robust_json --cov-report=html\n```\n\n### \ud83c\udfa8 Areas We'd Love Help With\n\n- **\ud83c\udf0d Internationalization**: Better support for non-Latin scripts and RTL languages\n- **\u26a1 Performance**: Optimize parsing speed for very large JSON objects\n- **\ud83d\udd0d LLM Integration**: Improve extraction from more LLM output formats\n- **\ud83d\udcda Documentation**: Examples, tutorials, and API documentation\n- **\ud83e\uddea Test Coverage**: Add more edge cases and real-world scenarios\n- **\ud83d\udc1b Bug Fixes**: Help us get to 100% test pass rate!\n\n### \ud83d\udccb Development Guidelines\n\n- **Code Style**: Follow PEP 8, use type hints, and add docstrings\n- **Testing**: Add tests for new features and bug fixes\n- **Documentation**: Update README and docstrings as needed\n- **Performance**: Consider performance impact of changes\n- **Compatibility**: Maintain Python 3.9+ compatibility\n\n### \ud83c\udfc6 Recognition\n\nContributors will be recognized in our README and release notes. We appreciate every contribution, no matter how small!\n\n**Ready to get started?** Check out our [open issues](https://github.com/callzhang/robust-json/issues) or start with the failing tests above!\n\n---\n\n## \ud83d\udcdd License\n\nThis project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.\n\n---\n\n## \ud83d\ude4f Acknowledgments\n\nBuilt for developers working with LLM-generated content who need reliability without sacrificing flexibility.\n\n---\n\n**Made with \u2764\ufe0f for the AI/LLM community**\n",
"bugtrack_url": null,
"license": null,
"summary": "Robust JSON extraction and repair utilities for LLM-generated content.",
"version": "0.1.5",
"project_urls": {
"Bug Tracker": "https://github.com/callzhang/robust-json/issues",
"Documentation": "https://github.com/callzhang/robust-json#readme",
"Homepage": "https://github.com/callzhang/robust-json",
"Repository": "https://github.com/callzhang/robust-json"
},
"split_keywords": [
"json",
" llm",
" ai",
" validation",
" repair",
" parser",
" extraction",
" chatgpt",
" claude"
],
"urls": [
{
"comment_text": null,
"digests": {
"blake2b_256": "04da328894a700a03dcb5acf2d355eb87fdc48bc0017e4ca779b9d2395d45344",
"md5": "e7020ce327ee7eac811f5e03ad4e0b5b",
"sha256": "607d377cadf6330f1399e6858d41251db4fce3307adc3aa8983eac6bd3c1bb49"
},
"downloads": -1,
"filename": "robust_json_parser-0.1.5-py3-none-any.whl",
"has_sig": false,
"md5_digest": "e7020ce327ee7eac811f5e03ad4e0b5b",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.9",
"size": 13259,
"upload_time": "2025-10-16T03:45:44",
"upload_time_iso_8601": "2025-10-16T03:45:44.050767Z",
"url": "https://files.pythonhosted.org/packages/04/da/328894a700a03dcb5acf2d355eb87fdc48bc0017e4ca779b9d2395d45344/robust_json_parser-0.1.5-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": null,
"digests": {
"blake2b_256": "7b45a210bf643e0a8123f2c7f2c142510c34194d959e95164bbb18c18318617c",
"md5": "791e74ddf8d8c8fb96c7100311c58aaf",
"sha256": "60a57ec6aba14fb7c9acffc745951cb1249dc35168841354af81571fcbab0193"
},
"downloads": -1,
"filename": "robust_json_parser-0.1.5.tar.gz",
"has_sig": false,
"md5_digest": "791e74ddf8d8c8fb96c7100311c58aaf",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.9",
"size": 32707,
"upload_time": "2025-10-16T03:45:45",
"upload_time_iso_8601": "2025-10-16T03:45:45.255487Z",
"url": "https://files.pythonhosted.org/packages/7b/45/a210bf643e0a8123f2c7f2c142510c34194d959e95164bbb18c18318617c/robust_json_parser-0.1.5.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2025-10-16 03:45:45",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "callzhang",
"github_project": "robust-json",
"travis_ci": false,
"coveralls": false,
"github_actions": false,
"lcname": "robust-json-parser"
}
JSON
