Skip to main content

a lib to parse llm answer to code

Project description

parse_llm_code

This library serves the purpose of parsing code snippets from text, particularly helpful when dealing with outputs generated by large language models. Typically, such outputs include both code snippets and explanatory text. By using this library, one can effectively filter and extract the meaningful code portions, aiding in the extraction of useful and actionable code from generated responses.

USECASE

Extract first code block from answer

from parse_llm_code import extract_first_code

test_string = """
this is llm answer example:

\```python
print('first line')
print('second line')
\```

 ```typescript
console.log('first line')
\```
"""

# Example of extracting the first code block
first_code = extract_first_code(test_string)
print(first_code.language)
print(first_code.context)
print(first_code.length)
print(first_code.lines)
print(first_code.to_dict())

Output

python

print('first line')
print('second line')

40

2

{'language': 'python', 'context': "print('first line')\nprint('second line')", 'length': 40, 'lines': 2, 'include_try': False, 'include_return': False}

Extract multi code blocks

from parse_llm_code import extract_code_blocks

test_string = """
this is llm answer example:

\```python
print('first line')
print('second line')
\```

 ```typescript
console.log('first line')
\```
"""

# Example of extracting all code blocks
result = extract_code_blocks(test_string)
print(result.length)
print(result.code_list)
print(result.code_dict_list)

Output

2

[<parse_llm_code.extract_code.CodeBlock object at 0x1049edac0>, <parse_llm_code.extract_code.CodeBlock object at 0x1049edfd0>]

[
    {
        "language": "python",
        "context": "print('first line')\nprint('second line')",
        "length": 40,
        "lines": 2,
        "include_try": False,
        "include_return": False,
    },
    {
        "language": "typescript",
        "context": "console.log('first line')",
        "length": 25,
        "lines": 1,
        "include_try": False,
        "include_return": False,
    },
]

Run test

python3 -m unittest discover -s tests

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

parse_llm_code-0.1.5.tar.gz (3.1 kB view details)

Uploaded Source

Built Distribution

parse_llm_code-0.1.5-py3-none-any.whl (4.3 kB view details)

Uploaded Python 3

File details

Details for the file parse_llm_code-0.1.5.tar.gz.

File metadata

  • Download URL: parse_llm_code-0.1.5.tar.gz
  • Upload date:
  • Size: 3.1 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/5.0.0 CPython/3.9.6

File hashes

Hashes for parse_llm_code-0.1.5.tar.gz
Algorithm Hash digest
SHA256 f2fad063329b5974a7cef13ce2ca728a3a946750172eba63f11c0a0c97f4e9f4
MD5 8c6a491a23512c612335112b944e3d5d
BLAKE2b-256 461f88c3f6ca2ad61f0ec88608aba44764240db7f46e606fa92df7517d1e204b

See more details on using hashes here.

File details

Details for the file parse_llm_code-0.1.5-py3-none-any.whl.

File metadata

File hashes

Hashes for parse_llm_code-0.1.5-py3-none-any.whl
Algorithm Hash digest
SHA256 78548d9a63670e9ddc19fd7030abf2fd063f54b04cc070b074cf9a4910839982
MD5 b8352615c044657ea4ae762be7f8e442
BLAKE2b-256 48d97b66854c8408e4372b779343927e4473f22a536fb6aa992246116874d434

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page