Incorrect encoding detection in Reports.get_report_document for ISO-8859-1 documents in version 2.1.5

[TSLdob]

Describe
In version 1.9.60, the Reports.get_report_document method correctly identified the encoding of report documents as ISO-8859-1. However, in version 2.1.5, the same documents are incorrectly detected as UTF-8, leading to potential issues with parsing or displaying content for certain locales and report types.

To Reproduce

Reports().get_report_document(
    reportDocumentId=reportDocumentId,
    download=True,
)

Observe the encoding detected in document_response.encoding for the document.

Expected behavior

The encoding should be correctly identified as ISO-8859-1 for documents from Marketplaces.JP with the report type "GET_LEDGER_DETAIL_VIEW_DATA".

Desktop:

• OS: Windows

Additional context

This issue appears to be specific to certain marketplaces and report types, such as Marketplaces.JP with the report type "GET_LEDGER_DETAIL_VIEW_DATA". Other marketplaces or report types may not be affected.

[saleweaver]

@codex check the difference between the two versions

[chatgpt-codex-connector]

Summary

• Compared behavior across versions and implemented a follow-up fix for the encoding regression introduced after the HTTP client migration: resolve_character_code now prefers explicit response charset metadata, and avoids trusting HTTPX’s default utf-8 guess when charset is missing (falls back to iso-8859-1 instead). sp_api/util/report_document.pyL6-L19
• Updated synchronous Reports.get_report_document to pass charset_encoding information into character-code resolution for both streaming and non-streaming download paths. sp_api/api/reports/reports.pyL377-L398
• Updated asynchronous Reports.get_report_document with the same charset-resolution behavior for parity with sync behavior. sp_api/asyncio/api/reports/reports.pyL368-L397
• Added focused tests validating charset precedence, fallback-from-default-utf8 behavior, and windows-31j -> cp932 mapping for report document decoding logic. tests/api/test_report_document_util.pyL1-L34

PR / Commit

• Committed on branch work with message: fix: preserve report document charset detection with httpx (commit 15f317d).
• Created a follow-up PR via make_pr: “Follow-up: fix report document encoding regression after httpx migration”.

Testing

• ✅ pytest -q tests/api/test_report_document_util.py
• ⚠️ pytest -q tests/api/test_report_document_util.py tests/api/reports/test_reports.py::test_get_report_document_n_decrypt (fails in this environment due to missing SP-API credentials, not due to code logic)

View task →

[saleweaver]

Hi @TSLdob

latest should fix that, can you please check?

Thanks for raising this!