feat: organized-pool context, archive introspection, matching config - #13
Merged
Conversation
… multi-select & preview
Track previously organized files in a persistent ledger so re-runs skip them
automatically. New files that are byte-identical to organized content are
flagged as duplicates. Claude results are cached by content hash: per-file
descriptions by file hash, grouping by the full set of file hashes plus
existing folder labels — re-running over an unchanged set sends zero tokens.
Group labels can now be nested paths ("Work/Acme Corp/Website Redesign")
that become real directory trees. Claude is prompted with existing folder
names from the ledger so new files route into matching groups instead of
near-duplicate ones.
TUI gains multi-select (Enter to mark, batch d/m/n) and a spacebar preview
modal for images and text files, reusing the DiffState modal pattern.
Folders containing `.spindel-ignore` are excluded from all scanning.
CLI additions: --no-ledger, --ledger <path>.
Add decode_image() helper that tries the image crate first, then falls back to ImageMagick `magick convert` for formats the crate can't handle (e.g. HEIC/HEIF). Replaces four inline decode blocks. Also clears preview state when no image path is available, and updates test model references from claude-sonnet-4-20250514 to claude-sonnet-4-6.
…toggles When grouping new files, the AI now receives rich descriptions (summaries + tags) from sample files in each existing group — not just bare folder names — so it can match accurately even with very few new files. ZIP archives are introspected: inner files are extracted to a temp dir, analyzed individually through the full AI pipeline, and synthesized into a single description. Respects configurable limits (max files, max file size per entry). Both features default to on and can be toggled via [matching] config section or CLI flags (--no-organized-context, --no-introspect-archives). New dependencies: zip 2, tempfile 3 (promoted from dev-only).
🔒 Upgrade RequiredPrivate repository analysis requires Pro or Enterprise. Upgrade: https://ecc.tools/pricing?plan=pro ECC Tools keeps the core app open, and puts private repos, team features, and enterprise controls behind paid tiers. |
All conflicts were in regions where this branch adds new code (organized-pool context, archive introspection, matching config) that main does not have. Kept our side in all cases.
🔒 Upgrade RequiredPrivate repository analysis requires Pro or Enterprise. Upgrade: https://ecc.tools/pricing?plan=pro ECC Tools keeps the core app open, and puts private repos, team features, and enterprise controls behind paid tiers. |
🔒 Upgrade RequiredPrivate repository analysis requires Pro or Enterprise. Upgrade: https://ecc.tools/pricing?plan=pro ECC Tools keeps the core app open, and puts private repos, team features, and enterprise controls behind paid tiers. |
zackkitzmiller
added a commit
that referenced
this pull request
Jun 16, 2026
…13) * feat: persistent organized ledger, grouping cache, nested labels, TUI multi-select & preview Track previously organized files in a persistent ledger so re-runs skip them automatically. New files that are byte-identical to organized content are flagged as duplicates. Claude results are cached by content hash: per-file descriptions by file hash, grouping by the full set of file hashes plus existing folder labels — re-running over an unchanged set sends zero tokens. Group labels can now be nested paths ("Work/Acme Corp/Website Redesign") that become real directory trees. Claude is prompted with existing folder names from the ledger so new files route into matching groups instead of near-duplicate ones. TUI gains multi-select (Enter to mark, batch d/m/n) and a spacebar preview modal for images and text files, reusing the DiffState modal pattern. Folders containing `.spindel-ignore` are excluded from all scanning. CLI additions: --no-ledger, --ledger <path>. * fix: HEIC/HEIF image preview via magick fallback, update test model refs Add decode_image() helper that tries the image crate first, then falls back to ImageMagick `magick convert` for formats the crate can't handle (e.g. HEIC/HEIF). Replaces four inline decode blocks. Also clears preview state when no image path is available, and updates test model references from claude-sonnet-4-20250514 to claude-sonnet-4-6. * feat: organized-pool context, archive introspection, matching config toggles When grouping new files, the AI now receives rich descriptions (summaries + tags) from sample files in each existing group — not just bare folder names — so it can match accurately even with very few new files. ZIP archives are introspected: inner files are extracted to a temp dir, analyzed individually through the full AI pipeline, and synthesized into a single description. Respects configurable limits (max files, max file size per entry). Both features default to on and can be toggled via [matching] config section or CLI flags (--no-organized-context, --no-introspect-archives). New dependencies: zip 2, tempfile 3 (promoted from dev-only). * gitginore
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
[matching]config section; disable with--no-organized-contextor--no-introspect-archivesCLI flagsChanges
src/config.rs— NewMatchingConfigstruct withuse_organized_context,introspect_archives,max_archive_files,max_archive_file_size_mb; CLI flags--no-organized-contextand--no-introspect-archivessrc/analyze/mod.rs—analyze_archive(),extract_and_analyze_archive(),extract_zip()for archive introspection with recursive analysis viaBox::pinsrc/ai/prompts.rs—group_organized_context()builds rich prompt section with sample file summaries/tags per groupsrc/ai/mod.rs—propose_groups_with_organized_context()trait method with default implsrc/ai/claude.rs—ClaudeProvideroverride using organized context in grouping promptsrc/ledger/mod.rs—group_content_hashes()returns blake3 hashes per group for cache lookupssrc/pipeline.rs—load_organized_context()loads cached descriptions; wired throughrun_ai_pipelinezip = "2",tempfile = "3"(promoted from dev-only)Test plan
cargo test)cargo clippycleancargo fmt --checkcleanMatchingConfigdefaults, TOML parsing, CLI flag overrides (7 new tests)group_content_hashesreturns/filters correctly (2 new tests)group_organized_contextempty/populated/truncation (3 new tests)extract_zipextracts files, skips hidden/nested, respects max limit (3 new tests)