Tech detector counts vendored and build-output files, skewing language detection

[ascherj]

tech_detector.py does not exclude node_modules/ or build/ paths. A repo with 2 Python source files and 6 vendored/bundled JS files is reported as primarily JavaScript.

Steps to reproduce:

from agent.tools.tech_detector import TechDetector
t = TechDetector()
files = ['main.py','core/app.py','node_modules/lib/index.js','node_modules/lib/util.js','node_modules/x/a.js','node_modules/y/b.js','build/bundle.js','build/vendor.js']
print(t.execute({'files': files}).data['primary_language'])
# observed: 'JavaScript'  (expected: 'Python')

Related failing tests: test_node_modules_excluded, test_build_directory_excluded in tests/unit/test_tech_detector.py.

[solo6760]

If nobody else has this I'll take this ticket

[fuentesjm]

I will be working on this. -fuentesjm

[paulshao2698]

I will work on this as well.

[MichaelHP23]

i will work on this

[VincentTLe]

I will working on this problem

[jadinee]

Hello, I selected this issue on the Google Sheet.

[GolamMortuzaSourov]

i will work on this

[GGirishya]

I like this, and i would love to work on it.

[Dennis-1am]

Hi I would like to work on this.