feat: add timestamp fields to picker API entries

[Cassin01]

Add created_at and updated_at to all API entry tables (get_memos, get_memos_by_tag, get_memos_for_link) for picker sorting support.

• parse_date_to_timestamp: extract epoch time from YYYYMMDD_HHMMSS filenames
• get_file_mtime: read file modification time via vim.uv.fs_stat
• Document Picker API in vim help (doc/sm.txt)

Summary

Brief description of the changes in this PR.

Related Issue

Fixes #

Type of Change

• Bug fix (non-breaking change that fixes an issue)
• New feature (non-breaking change that adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Documentation update

Checklist

• My code follows the project's code style (Fennel, not Lua directly)
• I have run make to regenerate Lua files
• I have run make test and all tests pass
• I have added tests for new functionality
• I have updated the documentation (if needed)

Testing

Describe how you tested your changes:

Screenshots (if applicable)

Add screenshots to help explain your changes.

[Copilot]

Pull request overview

This PR extends the picker-facing API entries with created_at and updated_at timestamps to support picker-side sorting, and updates tests/docs accordingly.

Changes:

• Add timestamp extraction (created_at from filename, updated_at from filesystem mtime) to memo metadata.
• Thread created_at/updated_at through picker API entry tables and update API/memo tests.
• Document the Picker API entry schema in doc/sm.txt.

Reviewed changes

Copilot reviewed 9 out of 10 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
fnl/sm/memo.fnl	Adds timestamp helpers and includes created_at/updated_at in get_memo_info.
fnl/sm/api.fnl	Adds timestamps to picker entry tables returned by API list functions.
fnl/sm/memo_test.fnl	Adds tests for date parsing + timestamp fields; stubs vim.uv.fs_stat.
fnl/sm/api_test.fnl	Updates API mocks/assertions to include timestamp fields.
lua/sm/memo.lua	Generated Lua reflecting new timestamp logic in memo module.
lua/sm/api.lua	Generated Lua reflecting new timestamp fields in API entries.
lua/sm/memo_test.lua	Generated Lua reflecting updated memo tests/mocks.
lua/sm/api_test.lua	Generated Lua reflecting updated API tests/mocks.
doc/sm.txt	Adds a “Picker API” help section documenting entry fields including timestamps.
.gitignore	Adds .macot (appears unrelated to PR goal).

[Copilot]

parse_date_to_timestamp will accept any 6 numeric components that match the pattern, but os.time normalizes out-of-range values (e.g. month=13, hour=99) instead of failing. That means clearly-invalid filenames could produce a non-nil created_at that is incorrect. Consider validating ranges (and ideally round-tripping with os.date to ensure the resulting fields match the parsed fields) and returning nil when the date/time is not a real YYYYMMDD_HHMMSS.

Suggested change

	(os.time {:year (tonumber year)
	:month (tonumber month)
	:day (tonumber day)
	:hour (tonumber hour)
	:min (tonumber min)
	:sec (tonumber sec)})))))
	(let [year-num (tonumber year)
	month-num (tonumber month)
	day-num (tonumber day)
	hour-num (tonumber hour)
	min-num (tonumber min)
	sec-num (tonumber sec)]
	;; Basic range validation to avoid obvious invalid dates/times
	(when (and year-num
	month-num (>= month-num 1) (<= month-num 12)
	day-num (>= day-num 1) (<= day-num 31)
	hour-num (>= hour-num 0) (<= hour-num 23)
	min-num (>= min-num 0) (<= min-num 59)
	sec-num (>= sec-num 0) (<= sec-num 59))
	(let [ts (os.time {:year year-num
	:month month-num
	:day day-num
	:hour hour-num
	:min min-num
	:sec sec-num})]
	;; Round-trip via os.date to ensure os.time didn't normalize
	(when ts
	(let [t (os.date "*t" ts)]
	(when (and (= t.year year-num)
	(= t.month month-num)
	(= t.day day-num)
	(= t.hour hour-num)
	(= t.min min-num)
	(= t.sec sec-num))
	ts)))))))))

[Copilot]

get_memo_info now calls uv.fs_stat (via get_file_mtime) for every memo, which adds a synchronous filesystem stat per entry. For large memo collections this can noticeably slow picker population. Consider making updated_at optional/lazy (computed in the picker integration layer only when needed), or caching stat results during a single listing call to avoid repeated stats.