fix: váha prvku komponent počítaná z komponent prošlých filtrem (#55) (#83)

* fix: váha prvku komponent počítaná z komponent prošlých filtrem (#55)

Váha prvku_vaha = 1/n se dosud počítala ze všech komponent dokumentační
jednotky ještě před filtrem na období a areál, takže při aktivním filtru
váhy prvků jedné DJ nedávaly v součtu 1 (DJ se 4 komponentami, filtru
vyhoví 1 → váha 0,25 místo 1).

- nová funkce _component_entries: nejdřív vyfiltruje komponenty, pak
  přidělí váhu 1/n z těch, které prošly; DJ bez komponent má váhu 1
- smoke test vaha_komponent (4×0,25; 1 ze 4 → 1; 2 ze 3 → 2×0,5; bez
  komponent → 1)
- README: pole prvek_vaha v tabulce atributů komponent
- changelog v2.2.0 doplněn (bez povýšení verze)
- OpenSpec změna openspec/changes/fix-component-feature-weight

Ověřeno: živá data (Praha, novověk) – 0 z 1204 DJ se součtem vah ≠ 1,
původní kód 921; check_sources, bandit, detect-secrets, flake8, ruff,
pyqgis4-checker, smoke test v qgis/qgis:ltr i :stable.

Implementace připravena AI (Claude, subagent), ověřena a zkontrolována.

* openspec: archivovat fix-component-feature-weight (#55)

Ruční test v QGIS ověřen správcem: akce i lokality s Načíst komponenty,
bez filtru i s filtrem období – váhy prvků jedné DJ dávají součet 1 a
počítají se jen z vyfiltrovaných komponent. Úkol 3.2 odškrtnut, změna
archivována přes openspec archive --skip-specs.

Připraveno s pomocí AI (Claude).

openspec/changes/archive/2026-10-02-fix-component-feature-weight/.openspec.yaml

@@ -0,0 +1,2 @@
+schema: spec-driven
+created: 2026-10-02

openspec/changes/archive/2026-10-02-fix-component-feature-weight/design.md

@@ -0,0 +1,49 @@
+## Context
+
+`load_amcr_data` in `amcr_viewer/amcr_tools.py` (section B) builds, for each
+DJ with a PIAN and *Načíst komponenty* on, one metadata dict per component
+and appends it to `pian_lookup[pian_id]`. The weight is set as
+`'vaha': 1/komps_count` with `komps_count = len(komps)` computed **before**
+the loop that skips components failing `komp_projde_filtrem`. The empty-DJ
+branch leaves `vaha` out and the feature builder falls back to
+`meta.get('vaha', 1)`.
+
+## Goals / Non-Goals
+
+**Goals:** weights of one DJ sum to 1 under any filter; the rule is
+testable offline.
+
+**Non-Goals:** weighting across DJs or across records that share one PIAN
+(a PIAN shared by several DJs still yields several features – the weight
+only de-duplicates components of one DJ, as #55 asked); changing the layer
+schema.
+
+## Decisions
+
+1. **Filter first, then weigh.** Build the list of passing components, then
+   set `vaha = 1/len(passing)`. Alternative – a second counting pass with
+   `sum(komp_projde_filtrem(...))` – duplicates the filter call and drifts
+   if the filter changes (#70 replaces it).
+2. **Extract a small pure helper** (e.g. `_component_entries(dj_meta, komps,
+   passes)` returning the list of per-component dicts with `vaha`, where
+   `passes` is a predicate) so the smoke test can check weights without
+   QGIS layers or network. The helper must not depend on how components are
+   selected, so #70 can pass a different predicate.
+3. **Explicit weight 1 for a DJ without components** instead of relying on
+   the `meta.get('vaha', 1)` default – the default stays as a safety net.
+
+## Risks / Trade-offs
+
+- Floating-point: 1/3 weights sum to 0.999…; acceptable for analyses,
+  test with a tolerance.
+- The helper extraction touches a long function; keep the diff limited to
+  the component branch.
+
+## Verification
+
+- Smoke test cases: 4 components no filter → 4×0.25; filter keeps 1 of 4 →
+  weight 1; keeps 2 of 3 → 2×0.5; no components → 1 entry, weight 1.
+- Full AGENTS.md check set (ltr + stable smoke test, pyqgis4-checker).
+- Manual QGIS test by the user: download akce with *Načíst komponenty* and
+  a period filter, check in the attribute table that `prvek_vaha` sums to 1
+  per `dj_id`.

openspec/changes/archive/2026-10-02-fix-component-feature-weight/proposal.md

@@ -0,0 +1,42 @@
+## Why
+
+Issue #55 added the `prvek_vaha` (feature weight) attribute: when *Načíst
+komponenty* is on, every component of a documentation unit (DJ) becomes its
+own feature on the same PIAN geometry, and the weight 1/*n* lets spatial
+analyses count the geometry once. The unreleased implementation on
+`version/v2.2.0` takes *n* from **all** components of the DJ, before the
+period/area filter. With a component filter active the weights of one DJ no
+longer sum to 1 (DJ with 4 components, 1 passes the Neolithic filter → one
+feature with weight 0.25 instead of 1), so weighted counts are wrong exactly
+when users filter. See the comment on #55.
+
+## What Changes
+
+- *n* in `prvek_vaha = 1/n` is the number of component features actually
+  created for the DJ, i.e. components that pass the period/area filters.
+- The weights of all features created from one DJ sum to 1 with or without
+  filters.
+- A DJ without components keeps its single feature with weight 1 (today the
+  value comes from a default; it becomes explicit).
+- `README.md` documents `prvek_vaha` in the component fields table (it is
+  missing there today).
+- Changelog entry under v2.2.0 in `amcr_viewer/metadata.txt` is extended
+  (the feature is unreleased, no separate version bump).
+
+## Capabilities
+
+### New Capabilities
+- `component-features`: one feature per component of a fieldwork event or
+  site, and the weight attribute that de-duplicates shared geometries.
+
+### Modified Capabilities
+
+## Impact
+
+- `amcr_viewer/amcr_tools.py` – component feature creation in
+  `load_amcr_data` (section B, attribute parsing).
+- `tests/smoke_test.py` – offline check of the weights.
+- `README.md`, `amcr_viewer/metadata.txt` (changelog only).
+- No change to the digiarchiv API contract, layer schema or stored settings.
+- `filter-components-via-component-endpoint` (#70) changes how components
+  are selected; it builds on this change and must keep the weight rule.

openspec/changes/archive/2026-10-02-fix-component-feature-weight/specs/component-features/spec.md

@@ -0,0 +1,35 @@
+# Spec Delta
+
+## Purpose
+
+Describes how components of fieldwork events and sites become map features
+and how their weight lets spatial analyses count a shared geometry once.
+
+## ADDED Requirements
+
+### Requirement: Weight of component features sums to one per DJ
+When components are loaded as features, each feature SHALL carry the weight
+`prvek_vaha = 1/n`, where *n* is the number of features created from the
+same documentation unit in this download. Components excluded by the period
+or area filter SHALL NOT count towards *n*.
+
+#### Scenario: No component filter
+- **WHEN** a documentation unit has 4 components and no period or area filter is set
+- **THEN** 4 features are created, each with weight 0.25
+
+#### Scenario: Filter keeps some components
+- **WHEN** a documentation unit has 4 components and the period filter matches 1 of them
+- **THEN** 1 feature is created with weight 1
+
+#### Scenario: Filter keeps two of three components
+- **WHEN** a documentation unit has 3 components and the filter matches 2 of them
+- **THEN** 2 features are created, each with weight 0.5, and their weights sum to 1
+
+### Requirement: Documentation unit without components has weight one
+When components are loaded and a documentation unit has no component, the
+single feature created for it SHALL have weight 1 and empty component
+fields.
+
+#### Scenario: DJ without components, no filter
+- **WHEN** a documentation unit with a PIAN has no components and no component filter is set
+- **THEN** one feature is created with empty component fields and weight 1

openspec/changes/archive/2026-10-02-fix-component-feature-weight/tasks.md

@@ -0,0 +1,37 @@
+# Tasks
+
+## 1. Weight computed from passing components
+
+- [x] 1.1 In `amcr_viewer/amcr_tools.py` extract the per-component entry
+  building of the *Načíst komponenty* branch into a pure helper that takes
+  the DJ metadata, the component list and a pass predicate, filters first
+  and sets `vaha = 1/len(passing)`; set `vaha = 1` explicitly for a DJ
+  without components; verify with `python3 tests/check_sources.py`,
+  `flake8 --isolated amcr_viewer/` and `ruff check .`
+- [x] 1.2 Extend `tests/smoke_test.py` with an offline case for the helper
+  (4 components no filter → 4×0.25; 1 of 4 passes → 1.0; 2 of 3 pass →
+  2×0.5, sum 1 within tolerance; no components → 1 entry, weight 1);
+  verify the smoke test passes in `qgis/qgis:ltr` and `qgis/qgis:stable`
+
+## 2. Documentation
+
+- [x] 2.1 Add `prvek_vaha` (alias *Váha prvku*) to the component fields
+  table in `README.md` with the rule "1/n, n = features created from the
+  same documentation unit after filters"; verify by reading the rendered
+  table
+- [x] 2.2 Extend the v2.2.0 changelog bullet about the feature weight in
+  `amcr_viewer/metadata.txt` (weights of one documentation unit sum to 1
+  also with period/area filters); no version bump – 2.2.0 is unreleased and
+  `CITATION.cff` already says 2.2.0; verify both versions match
+
+## 3. Verification
+
+- [x] 3.1 Run the full local check set from `AGENTS.md` (check_sources,
+  bandit, detect-secrets `--all-files`, flake8 `--isolated`, ruff,
+  pyqgis4-checker log empty, smoke test ltr + stable); verify all clean
+- [x] 3.2 Manual test in QGIS (user): akce in a small window with *Načíst
+  komponenty* and one period filter; verify in the attribute table that
+  `prvek_vaha` sums to 1 per `dj_id`
+  - Verified by the maintainer 2026-10-02: akce and lokality with
+    *Načíst komponenty*, without and with a period filter – weights sum
+    to 1 per DJ and are computed only from the filtered components