Repository navigation
Releases
1.9.10
Automatic Mood History 1.9.10
- Composition over time shows up to ten of the most common emotions in the selected date range. Shorter or custom ranges can surface emotions that are uncommon across the full history.
- Colors follow the emotion across date ranges, with gold for joy, red for anger, blue for sadness, and a consistent color for each custom label. The legend, chart, and Common Emotions bars use the same assignments.
This display update uses your existing mood history. No reanalysis is needed.
1.9.9
Automatic Mood History 1.9.9
- Keep saved detected-date results visible in the chart and daily list after source membership changes.
- Refresh an unedited monthly export that contains an exact subset of saved records. Preserve edited exports while committing new results to history immediately.
- Accept explicit Groq assessments of insufficient emotional evidence without requiring a valid numerical score or scored valence. Scored assessments still undergo full validation.
1.9.8
1.9.8
Future cache verification timestamps could bypass the daily content check after clock rollback or sync. Revalidate immediately when cache age is negative. Overlarge Retry-After values could throw during rate-limit handling. Fall back to the standard pause for unrepresentable deadlines.
1.9.7
Trend presets use binary search for the date boundary. History reuses a month formatter and avoids a redundant sort. Month counts, dashboard typography and the primary analysis action are easier to scan.
1.9.5
Automatic Mood History 1.9.5
This release reduces work when opening and refreshing a large mood history.
- Day-by-day records are grouped under collapsed month headings. Entry counts stay visible, and detailed score cards are created when a month is expanded. Existing record details and source links remain available.
- Filename date parsing is cached in a bounded cache keyed by the filename and relevant date settings. Built-in date-format parsing is skipped for ordinary undated titles, while a configured Daily Notes format is still checked.
- Opening an existing Mood History dashboard no longer checks or rewrites support files.
Scores, saved history, and source notes are unchanged. Native Obsidian performance has not yet been measured on a personal vault.
1.9.4
Composition chart spacing
The composition chart now adds a small gap between daily bars in the 7?90 day and custom ranges, using the same spacing rule as the 180-day chart.
1.9.3
Recover saved history without repeated analysis
Automatic Mood History now completes interrupted monthly exports in batches. It restores existing history from monthly notes first, then saves the already-computed results without calling Groq again. This addresses large backlogs that could make Obsidian lag on startup.
If a monthly note contains personal edits, the plugin keeps that note intact and makes the computed scores available in its JSON history. The dashboard shows an export warning. To refresh that month, move the edited monthly note aside and choose Export months.
Ordinary scans skip unchanged note reads when file metadata is unchanged. A source event triggers an immediate check, and a full content check still runs at least daily to catch sync changes. Saved fallback scores are not upgraded by ordinary scans unless a retry is already pending; Retry fallback entries and Retry all with Groq remain available for explicit upgrades.
Automatic recovery handles at most five pending provider retries per pass, then continues on later passes. This keeps a large older backlog from monopolizing Obsidian.
The Morning Pages word-list algorithm is the offline fallback. The separate Moodring model comparison has been removed from the plugin, settings, release assets, and build. Existing saved comparison sidecars remain in historical records but are no longer run or displayed.
The composition chart's daily columns now fill their slots across the 7–90 day ranges, eliminating gaps between adjacent bars. Both Automatic Mood History and Morning Pages Analytics offer an opening range preference. Choose Last used to reopen the previous preset or custom range; an explicit default takes precedence on opening.
1.9.2
Restore the chart from monthly notes
When saved history is empty, Automatic Mood History now restores records from the recovery data in its monthly notes at startup. It redraws the dashboard using the saved scores. Recovery doesn't run analysis or call Groq.
Use Restore from monthly notes on the empty dashboard or Automatic Mood History: Restore history from monthly notes in the command palette to recover missing records after sync. Restore preserves existing scores and leaves monthly notes intact.
Monthly exports are also protected against writes that would discard saved records missing from the plugin's data file.
1.9.1
Retry and threshold recovery fixes
When a Groq retry fails, Automatic Mood History now preserves any unchanged complete saved score regardless of which analyzer produced it. New or changed notes remain eligible for retry under the existing provider backoff behavior.
Lowering Minimum characters now reevaluates notes previously marked insufficient even when Groq is enabled, allowing eligible writing to be analyzed with Groq.
The regression tests use synthetic notes and a stubbed provider. Functional, mobile runtime, desktop/mobile browser layout and release privacy checks passed. Native Obsidian behavior and live Groq requests have not been verified for this release.
1.9.0
Morning Pages fallback and dashboard updates
Automatic Mood History now uses bundled Morning Pages Analytics feelings and topics word lists to score new or changed writing when Groq is disabled or unavailable. Existing Groq and older local scores remain unchanged for unchanged writing. Groq retry backoff remains in place, and saved fallback results can be retried with Groq.
Fallback scores have low confidence. The bundled matcher uses built-in word lists, does not use another plugin's custom lists, and does not account for negation, quotations, or who a feeling belongs to. It is a lightweight estimate, not a contextual reading.
The dashboard now offers a 180-day trend range. Selected preset and custom ranges survive automatic redraws, while a newly opened dashboard still starts at 14 days. A dashboard remains registered through temporary DOM detachment until its Obsidian component unloads. Composition charts show gray columns for every period without a top-ten emotion and period details distinguish an empty period from scored records with no displayed emotion.
The task wording “Grok” was interpreted as Groq, the provider already used by this plugin. This release does not add xAI Grok support.
Automated checks use synthetic notes and a stubbed provider. Browser layout, native Obsidian behavior, and a live Groq request were not verified for this release.