Replace requests with urllib.request to reduce core dependencies

[Copilot]

What does this changes

Replaces the requests library with Python's standard library urllib.request for HTTP operations, reducing core dependencies from 2 to 1 (only tzdata for Windows remains).

What was wrong

PyThaiNLP had requests>=2.31 as a core dependency solely for corpus download functionality (get_corpus_db() and _download() in pythainlp/corpus/core.py). This contributed to:

• Long dependency installation times (>1 hour per platform)
• Increased risk of subdependency version conflicts
• Larger installation footprint

How this fixes it

• pythainlp/corpus/core.py: Replaced requests.get() with urllib.request.urlopen()
  • Created _ResponseWrapper class for API compatibility with existing .json() method calls
  • Maintained chunked streaming downloads using response.read(CHUNK_SIZE)
  • Updated exception handling from requests.exceptions.* to urllib.error.*
  • Added library-appropriate User-Agent string that identifies PyThaiNLP client with version and platform info (e.g., PyThaiNLP/5.2.0 (Python/3.12; linux))
  • Defined _USER_AGENT module-level constant used consistently across HTTP requests
• pyproject.toml: Removed requests>=2.31 from core dependencies
• CHANGELOG.md: Documented the change

No breaking changes—API surface remains identical.

Your checklist for this pull request

• Passed code styles and structures
• Passed code linting checks and unit test

Original prompt

---

This section details on the original issue you should resolve

<issue_title>Consider reduce dependencies</issue_title>
<issue_description>## Background

• It is becoming more difficult to build the entire PyThaiNLP package since it has a long list of dependencies, and each of them may require conflicting versions of subdependencies.
• Dependency installation alone can take longer than one hour for each platform, way unhealthy to get responses for development
• Some dependencies also only support older/end-of-life versions of Python, making the situation more complicated. Some CI environments may not support these old Python versions. It is of security concerns too. See Python end of life information here https://devguide.python.org/versions/

Proposal

• Use the standard library when it possible. Often times there are functionalities that good enough for the job.
• For the features PyThaiNLP develops on our own, try to stick with a single library/framework for one particular area of purposes (for example, use either Tensorflow or PyTorch)
• Before adding a new dependency, assess first what are the transitive dependencies
• Lazy import when possible. (This does not help shortening the test time though)
• Consider putting the dependencies as a separate section in the optional dependencies in pyproject.toml. Keep the dependencies for the standard installation minimal. (This does not help shortening the test time though)

See also

• Porting model to ONNX model #639
• Remove all python-crfsuite models from PyThaiNLP #655

Notes

• ONNX Runtime packages will stop supporting Python 3.8 and Python 3.9 https://github.com/microsoft/onnxruntime/releases/tag/v1.20.0</issue_description>

<agent_instructions>Try to reduce the number of dependencies gradually.

Note that the unittest.yml workflow only test core and compact sets of features.
They are not the full features available in the package. You need to do additional testing for those features that are not covered in unittest.yml when you adjust the depedencies.</agent_instructions>

Comments on the Issue (you are @copilot in this section)

@wannaphong - Unnecessary base dependencies? #1184 - Fix base dependencies #1185

• Will improve Consider reduce dependencies #935

[github-actions]

Hello @Copilot, thank you for submitting a PR! We will respond as soon as possible.

สวัสดี @Copilot ขอบคุณที่ส่ง PR เข้ามา เราจะตอบกลับให้เร็วที่สุดเท่าที่จะทำได้

[bact]

@copilot considering our use case, what is the best user-agent string?

[Copilot]

@copilot considering our use case, what is the best user-agent string?

For this use case (a Python library downloading corpus data from its own server), the best User-Agent identifies the actual client rather than masquerading as a browser. Updated in commit 9507e0d to use: PyThaiNLP/{version} (Python/{py_version}; {platform})

This approach provides:

• Transparency in server logs
• Better debugging and analytics
• Follows conventions used by other Python libraries (pip, requests, etc.)
• Example output: PyThaiNLP/5.2.0 (Python/3.12; linux)

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud