ranaroussi/yfinance v1.2.0 Update #

ranaroussi/yfinance has been updated to v1.2.0. Minor updates seem to be occurring quite frequently since the v1.0.0 release.

Progress up to v1.1.0 #

• 2025-12-18: Project started — Goal: Fully port python-yfinance to Go.
• 2025-12-23: v1.0.0 Tag Release (Core packages: Ticker, Multi, Screener, Market, Sector, Industry, Calendars, etc.).
• 2026-01-26: v1.1.0 Release — Completed 5 phases of the Price Repair system.
  • Phase 1: Statistical utilities (percentile, zscore, filter).
  • Phase 2: Capital Gains double-counting fix + Stock Split repair.
  • Phase 3: 100x currency unit error fix + zero-value interpolation.
  • Phase 4: Fixes for 7 types of dividend errors.
  • Phase 5: ISO 10383 MIC code mapping for 60+ exchanges.
  • Stats: ~5,400 lines of code, 96 tests (100% pass).

v1.2.0 Development History #

1. Analysis of python-yfinance v1.2.0 Changes #

I analyzed the changes in python-yfinance v1.2.0 to see what needed to be reflected in Go. There were 4 non-screener changes, all of which were deemed unnecessary for the Go implementation:

• df._consolidate(): Go slices use contiguous memory; no pandas block fragmentation issues.
• flags.writeable guard: The Go repair package already performs defensive copying (make+copy).
• Typo fix (recieved→received): This typo did not exist in go-yfinance.
• version.py update: go-yfinance manages versions via git tags.

2. Identifying Missing Features in the Screener Module #

While analyzing the screener exchange map expansion for v1.2.0, I discovered that several core screener features that should have been in v1.1.0 were entirely missing. Initially, an AI tool suggested that no action was needed because the exchange maps weren’t in Go yet. However, realizing this contradicted the goal of exact functional parity, I pushed for a deeper analysis.

This experience highlighted the importance of human oversight. Even when porting a service solo with AI assistance, gaps can occur. It served as a reminder that AI should be treated as a “collaborative peer” rather than a tool for 100% delegation.

3. Implementation: feature/screener-parity #

I created a dedicated branch and implemented the following:

New Files (3)

• pkg/models/screener_const.go: Ported all constants from Python’s const.py.
  • Mapped all sectors, industries, regions, and peer groups.
  • Intentionally reproduced a Python bug (missing comma in income_statement) for compatibility.
• pkg/models/screener_query.go: Added EquityQuery/FundQuery types and ScreenerQueryBuilder interface.
  • Implemented ToDict() and QuoteType() methods.
  • Added logic for IS-IN expansion and field validation.
• pkg/screener/predefined.go: Implemented 15 predefined queries (9 Equity, 6 Fund).
  • Configured with fail-fast logic using init().

Updated Files (2)

• pkg/models/screener.go: Added fund-specific fields and OpISIN constants.
• pkg/screener/screener.go: Integrated the new interface and implemented a dynamic quoteType logic.

4. Discovery: Hidden Yahoo Finance API Bug #

During implementation, I found that the GET predefined endpoint silently ignores the offset parameter. To support pagination, I implemented a workaround in the Screen() method that switches to the POST custom query endpoint when offset > 0.

5. Testing and Finalization #

• Added 11 new test functions covering query generation, validation, and expansion.
• All 17 packages passed tests; golangci-lint reported 0 issues.
• Fully updated docs/API.md.
• 2026-02-19: Merged into main and deployed documentation.

Installation and Update #

go get https://github.com/wnjoon/go-yfinance@v1.2.0