v0.2.0: standalone viewer as app, narratology runtime, rename tvplotlines → tvplot#1
Merged
BirdInTheTree merged 24 commits intomainfrom Apr 15, 2026
Merged
Conversation
Build script assembles shell.html + parts (CSS, JS, demo data) into a single self-contained HTML file. Three-screen layout: welcome, grid, analytics with modal container and placeholder slots. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Keep battle-tested styles: color palette (light + Catppuccin Mocha dark), function colors, rank badges, tension variables, grid layout, analytics. Remove HTMX/edit-mode/merge/split styles (v1 is read-only). Add screen routing, welcome screen, and toolbar/navigation styles. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Port plotline x episode grid from tvplotlines-app Jinja2 template to vanilla JS. Includes event grid lookup, character resolution, guest detection, rank sorting, function legend filtering, season/character filters, and read-only event detail modal. App.js handles screen routing, dark mode persistence, demo data loading, and file upload. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Port computation (analytics.py) and rendering (analytics.html) to standalone JS: scorecard, arc map, episode pulse, convergence moments, character weight. Wire up Analytics button in app.js. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Store object centralizes all localStorage access (API keys, results, settings, onboarding state). Toolbar now has series dropdown, Grid/Analytics tab buttons, dark mode toggle, load/export/LLM/onboarding placeholders. Series switching re-renders current view and persists last selection. Drop zone overlay supports drag-and-drop and file picker for JSON import. Init logic checks onboarding state and restores last session. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Export dropdown in toolbar with 4 formats. Single floating menu shared between grid and analytics screens. CSV chosen over XLS to avoid SheetJS dependency — includes both events and plotlines summary. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Implement pipeline.js with streaming API client (Anthropic + OpenAI), all 5 passes with embedded prompts, post-processing, and synopsis file upload. Wire drop zone to accept .txt synopses alongside JSON, add LLM settings dialog, and pipeline progress overlay. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…ode labels - Force color-scheme to prevent Firefox dark mode override - .hidden !important to fix drop zone appearing on load - inciting_incident: no background, bright green text - crisis function color added - Story engine / logline moved above all analytics sections - Episode numbers shown once in scorecard header - Episode Pulse: removed themes, added plotline color legend Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Generate a standalone HTML viewer alongside the result JSON when --html is passed. Uses build_html() from the html module. Defaults to saving next to the JSON with .html extension; --html-output overrides the path. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Japanese minimalism aesthetic: lightweight typography, pill-shaped "Play me" button, subtle "Skip" link. First visit shows welcome, returning visits skip straight to grid. Smooth fade-out transition. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Animated walkthrough plays on actual app screens using Breaking Bad demo data: LLM settings modal, file upload, pipeline progress, grid reveal, analytics tour, and export. Ends with real LLM modal so user can connect their key. Skip link aborts animation cleanly. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…tem parameter - Rename prompts_en/ to prompts/hollywood/ - Import 8 narratology prompts (formerly tvplotlines_narratology repo) into prompts/narratology/ - Drop prompts_ru/ entirely — Russian support out of current scope - Replace lang parameter with system across library: get_plotlines, LLMConfig, CLI - Narratology pipeline raises NotImplementedError for now; prompts in place for v2 - Copy schema.md into docs/layered-schema.md (base + layers JSON design) - Update pyproject.toml artifacts, example, tests Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Pass 2: every event gets id {episode}#{index:02d} and keeps it downstream
- Passes 4–6: reference events by id, not by description
- Pass 5 mre: event_id instead of episode+description
- glossary: plain-English actant names (who_chases, bigger_force, …)
now mapped to their Greimas terms (subject, power, …)
- docs/layered-schema.md aligned with prompt output
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Welcome screen now has an "Analyze a series" form (show + season). Clicking Analyze asks the configured LLM for episode synopses from its training memory, shows them in an editable preview modal, then runs the 5-pass pipeline on whatever the user confirms. - LLM settings modal gains an "Analysis system" selector. Hollywood is live; Narratology is disabled with a "coming soon" note until its runtime lands in v2. - Exports: added "PDF (via print)" that calls window.print(); a print stylesheet hides the toolbar/modals and lays the grid out in landscape, so Cmd+P → "Save as PDF" produces a usable document. - Drag-and-drop .txt flow unchanged — it's the fallback when the model doesn't know the show. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…owledge move
1. Narratology pipeline is now runnable, Python and browser both:
- src/tvplotlines/narratology.py runs all six passes (context → fabula →
actants → story → arc → review) and maps results into the existing
TVPlotlinesResult shape so the viewer renders either system unchanged
- pipeline.js grows runPipelineNarratology mirroring the Python orchestrator
(per-episode and per-plotline passes parallelised), and the top-level
runPipeline dispatches on the user's system choice
- The "Narratology" option in LLM settings is enabled
2. Viewer prompts are no longer hardcoded. build.py reads every
prompts/{hollywood,narratology}/*.md at build time and injects them as a
_PROMPTS object, eliminating drift between source prompts and the viewer
3. Native PDF export — jsPDF is lazy-loaded from CDN and generates a vector
PDF from the loaded result (title, per-plotline sections with actants/
story-DNA + event list, episode appendix). Falls back to window.print()
when offline
4. Knowledge consolidated from 3-resources/tvplotlines/ into
docs/knowledge/ (decisions, specs, theory, prompting, results,
v3_narratology evergreens, archive). Personal artifacts stay in place
All 115 tests pass.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- docs/knowledge/ — moved to 3-resources/tvplotlines/. Knowledge,
evergreens, autoresearch drafts, and strategy notes belong out of the
public repo. Personal artifacts already lived there
- docs/viewer/{spec,plan,decisions}.md — internal Russian planning docs,
moved to 3-resources/tvplotlines/viewer-planning/
- .gitignore — exclude Untitled.md draft and the dacomply compliance/
scan output so they can sit in the working tree without leaking
- README — rewrite to reflect the current code: hollywood + narratology
systems with the --system flag, standalone HTML viewer mode, the
"Analyze a series" entry flow, --html/--html-output, narratology's
six-pass table, fixed the broken license badge and the dead
prompts_en/glossary.md link
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- api.md: get_plotlines signature gains `system`, `skip_review`, `base_url`, `suggested_plotlines`; new system-comparison table; Plotline gets reviewed_rank/computed_rank/nature/confidence, Event gets plot_fn; full CLI flag table for run + write-synopses - quickstart.md: --system and --html flags, standalone-viewer section with "Analyze a series" flow, system switch in Python example - index.md: mentions standalone viewer + layered-schema link - formulas.md: Pass 4 section (was missing), narratology pipeline section at the end documenting all six passes and the actant ↔ hero/goal/obstacle/stakes mapping Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Untitled.md and compliance/ live in 3-resources/tvplotlines/ alongside the rest of the private notes; they don't need to be ignored here any more. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Address the compliance scan items that actually apply to an installable library with no hosted service: - docs/ai-disclosure.md — single source of truth for data flow, vendor policies, and user rights. Alpine Animation (Switzerland) is named as publisher; we collect nothing. Names the three synopsis paths (user-supplied, Wikipedia-scraped via write-synopses, and the browser's "Analyze a series" flow), covers Anthropic AUP and OpenAI Usage Policies compliance, and records a use-case paper trail - docs/methodology.md — plain-language Art. 50 explainer written from the code (pipeline.py, narratology.py, pass*.py, postprocess.py), not the dacomply template. Points to formulas.md for line-by-line rules - README gains a "How this uses AI" section linking both docs - HTML viewer shows a dismissible AI-assisted-content footer naming the model family and Alpine Animation; dismissal persists in localStorage; footer is hidden at print time so PDF exports stay clean Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Alongside GitHub issues — gives data-subject and methodology requests a direct channel to the publisher, which is what the compliance scan flagged as missing. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replace <411@alpineanimation.ch> with "411 [at] alpineanimation [dot] ch". Defeats ~80–90% of off-the-shelf email harvesters that look for mailto: links or @-symbols; humans read it without effort. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
tvplotlines@alpineanimation.ch — easy to kill and recreate if it ever gets scraped, which is a better strategy than trying to obfuscate a shared catch-all mailbox. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Shorter, snappier. PyPI name, import path, CLI command (tvplot run), env vars (TVPLOT_*), localStorage keys, and all docs move in one step. Version bumps to 0.2.0 and CHANGELOG carries a BREAKING migration note. What changes externally: - pip install tvplot - from tvplot import get_plotlines - tvplot run my-show/ --system narratology - TVPLOT_OUTPUT_DIR, TVPLOT_SYNOPSES_DIR What stays: - Public API surface (get_plotlines signature, Plotline/Event models) - All pipeline logic and prompts - Existing result JSONs remain valid Rationale: no meaningful users yet on PyPI (80%+ of the 0.1.0 download count is bots and scanners), the rename is much cheaper now than after promotion. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Ten commits; the headline changes:
tvplotlines→tvplot(version bump to 0.2.0). PyPI name, import path, CLI command, env vars, and localStorage keys all move together. Old `tvplotlines 0.1.0` on PyPI will be yanked.Test plan
🤖 Generated with Claude Code