如何仅获取Pylint评分值？正则提取是否为唯一方案？

Great question! I’ve wrestled with this exact Pylint limitation before, so I feel your pain.

First, let’s confirm your findings: you’re 100% correct—Pylint doesn’t have a dedicated command-line flag or a straightforward method in epylint.py_run that returns only the numeric score (e.g., 7.5 or 7.5/10). All built-in output modes either dump the full report, or focus on errors/warnings without isolating the final score.

That said, regex extraction from standard output is the most common and straightforward solution, but it’s not strictly the only feasible approach. Here’s a breakdown of your options:

Option 1: Regex Extraction (Quick & Dirty) #

Pylint’s final score line follows a consistent format like:

Your code has been rated at 7.50/10 (previous run: 7.20/10, +0.30)

You can easily parse this with regex, either via command-line tools or in Python.

Command-Line Example #

pylint <package-to-scan> | grep -oP 'rated at \K[0-9.]+/[0-9]+'
# Or to get just the numeric value before the slash:
pylint <package-to-scan> | grep -oP 'rated at \K[0-9.]+'

Python Example (Using Subprocess) #

import subprocess
import re

result = subprocess.run(
    ["pylint", "<package-to-scan>"],
    capture_output=True,
    text=True
)

score_match = re.search(r"rated at ([0-9.]+)/[0-9]+", result.stdout)
if score_match:
    score = float(score_match.group(1))
    print(f"Pylint Score: {score}")

Option 2: Custom Pylint Reporter (Robust, Programmatic) #

If you want a more integrated, reliable solution (and don’t mind diving into Pylint’s API), you can create a custom reporter that only tracks and outputs the final score. This avoids relying on parsing stdout, which could break if Pylint changes its output format.

Here’s a minimal example:

from pylint.lint import Run
from pylint.reporters.base_reporter import BaseReporter
import re

class ScoreOnlyReporter(BaseReporter):
    def __init__(self, output=None):
        super().__init__(output)
        self.score = None

    def display_reports(self, layout):
        # Extract the score from the global stats section
        for section in layout.children:
            if hasattr(section, 'children'):
                for child in section.children:
                    if hasattr(child, 'data') and 'global_note' in child.data:
                        note_text = child.data['global_note']
                        score_match = re.search(r"rated at ([0-9.]+)/[0-9]+", note_text)
                        if score_match:
                            self.score = float(score_match.group(1))

# Run Pylint with the custom reporter
reporter = ScoreOnlyReporter()
Run(["<package-to-scan>"], reporter=reporter)
print(f"Pylint Score: {reporter.score}")

Verdict #

Regex extraction is the easiest and most widely used approach for most cases. The custom reporter is better if you’re building a tool that relies on Pylint’s score long-term, as it’s less fragile to output changes.

So to answer your question: no, regex extraction isn’t the only feasible solution, but it’s the most practical one for most scenarios.

内容的提问来源于stack exchange，提问作者Dev