davebb/linear-coding-agent
1
0
Fork 0
Code Issues Pull Requests 1 Actions Packages Projects Releases Wiki Activity

fix: Adapt hierarchical display for mismatched sectionPath formats

Browse Source 
Root cause:
- Summary.sectionPath: '635. As for the subject...' (paragraph numbers)
- Chunk.sectionPath: 'Peirce: CP 4.47 > 47. §3 THE NATURE...' (canonical refs)
- No way to match them with prefix/equal filters

Solution (workaround until summaries are regenerated):
- Show sections as **context** (relevant high-level topics found)
- Show chunks **globally** (top 20 most relevant passages)
- Don't try to group chunks under sections

UI changes:
- '📚 Sections pertinentes trouvées' (context cards with summary)
- '📄 Passages les plus pertinents' (top chunks, not grouped)
- Cleaner, more honest representation of what we found

Next steps to fully fix:
- Regenerate Summary collection with correct sectionPath format
- Or create a mapping between Summary titles and Chunk sectionPaths
This commit is contained in:
David Blanc Brioir
2026-01-01 15:51:11 +01:00
parent 47cf21867f
commit d824269606
2 changed files with 63 additions and 67 deletions
Download Patch File Download Diff File Expand all files Collapse all files

26
generations/library_rag/flask_app.py
View File

@@ -457,22 +457,18 @@ def hierarchical_search(
for obj in chunks_result.objects
]
# Distribute chunks to sections using prefix matching
all_chunks = []
# NOTE: Summary.sectionPath format doesn't match Chunk.sectionPath
# This is a data quality issue that needs to be fixed at ingestion
# For now, sections provide context, chunks are shown globally
print(f"[HIERARCHICAL] Got {len(all_chunks_list)} chunks total")
print(f"[HIERARCHICAL] Found {len(sections_data)} relevant sections")
all_chunks = all_chunks_list
# Clear chunks from sections (they're displayed separately)
for section in sections_data:
section_ref = section["section_path"] # e.g., "Peirce: CP 2.504"
# Find chunks whose sectionPath starts with this reference
section_chunks = [
chunk for chunk in all_chunks_list
if chunk.get("sectionPath", "").startswith(section_ref)
]
# Sort by similarity and limit per section
section_chunks.sort(key=lambda x: x.get("similarity", 0) or 0, reverse=True)
section["chunks"] = section_chunks[:limit]
section["chunks_count"] = len(section["chunks"])
all_chunks.extend(section["chunks"])
section["chunks"] = []
section["chunks_count"] = 0
# Sort all chunks by similarity (descending)
all_chunks.sort(key=lambda x: x.get("similarity", 0) or 0, reverse=True)

40
generations/library_rag/templates/search.html
View File

@@ -174,33 +174,35 @@
<!-- Hierarchical display -->
{% if results_data.mode == "hierarchical" and results_data.sections %}
<!-- Show relevant sections as context -->
<div style="margin-bottom: 2rem;">
<h3 style="font-size: 1.2em; margin-bottom: 1rem; color: var(--color-accent);">📚 Sections pertinentes trouvées</h3>
<div style="display: flex; flex-direction: column; gap: 1rem;">
{% for section in results_data.sections %}
<div class="section-group">
<div class="section-header">
<h3 style="margin: 0 0 0.5rem 0; font-size: 1.3em;">
📂 {{ section.title[:80] }}{% if section.title|length > 80 %}...{% endif %}
<span class="badge" style="background-color: var(--color-accent-alt); color: white; margin-left: 0.5rem;">{{ section.chunks_count }} passage{% if section.chunks_count > 1 %}s{% endif %}</span>
<span class="badge badge-similarity" style="margin-left: 0.5rem;">⚡ {{ section.similarity }}% similaire</span>
</h3>
{% if section.section_path and section.section_path != section.title %}
<p class="text-muted" style="margin: 0.25rem 0; font-size: 0.9em;">📍 {{ section.section_path }}</p>
{% endif %}
<div style="padding: 1rem; border-left: 3px solid var(--color-accent); background: rgba(125, 110, 88, 0.05); border-radius: 4px;">
<div style="display: flex; align-items: center; gap: 0.5rem; margin-bottom: 0.5rem;">
<strong>{{ section.title[:80] }}{% if section.title|length > 80 %}...{% endif %}</strong>
<span class="badge badge-similarity">⚡ {{ section.similarity }}% similaire</span>
</div>
{% if section.summary_text %}
<p class="summary-text" style="margin: 0.5rem 0; font-style: italic; color: #555;">{{ section.summary_text }}</p>
<p style="margin: 0; font-size: 0.9em; color: var(--color-text-main); font-style: italic;">{{ section.summary_text[:150] }}{% if section.summary_text|length > 150 %}...{% endif %}</p>
{% endif %}
{% if section.concepts %}
<div class="concepts" style="margin-top: 0.5rem;">
<div style="margin-top: 0.5rem;">
{% for concept in section.concepts %}
<span class="badge" style="background-color: rgba(125, 110, 88, 0.15); color: var(--color-accent); border: 1px solid rgba(125, 110, 88, 0.3); margin-right: 0.25rem;">{{ concept }}</span>
<span class="badge" style="background-color: rgba(125, 110, 88, 0.15); color: var(--color-accent); border: 1px solid rgba(125, 110, 88, 0.3); margin-right: 0.25rem; font-size: 0.85em;">{{ concept }}</span>
{% endfor %}
</div>
{% endif %}
</div>
{% endfor %}
</div>
</div>
<!-- Chunks within this section -->
{% if section.chunks %}
<div class="chunks-list" style="margin-left: 1.5rem; margin-top: 1rem;">
{% for chunk in section.chunks %}
<!-- Show top relevant chunks globally -->
<div>
<h3 style="font-size: 1.2em; margin-bottom: 1rem; color: var(--color-accent);">📄 Passages les plus pertinents</h3>
{% for chunk in results_data.results[:20] %}
<div class="chunk-item" style="background: white; padding: 1rem; margin-bottom: 0.75rem; border-left: 3px solid var(--color-accent-alt); border-radius: 4px;">
<div style="margin-bottom: 0.5rem; display: flex; gap: 0.5rem; flex-wrap: wrap; align-items: center;">
{% if chunk.work and chunk.work.author %}
@@ -213,7 +215,7 @@
</div>
<div class="passage-text" style="margin-bottom: 0.5rem;">"{{ chunk.text }}"</div>
<div class="passage-meta" style="font-size: 0.85em; color: #666;">
<strong>Section :</strong> {{ chunk.sectionPath or section.section_path or '—' }} &nbsp;│&nbsp;
<strong>Section :</strong> {{ chunk.sectionPath or '—' }} &nbsp;│&nbsp;
<strong>Type :</strong> {{ chunk.unitType or '—' }} &nbsp;│&nbsp;
<strong>Langue :</strong> {{ (chunk.language or '—') | upper }}
</div>
@@ -228,8 +230,6 @@
{% endfor %}
</div>
{% endif %}
</div>
{% endfor %}
<!-- Simple display (original) -->
{% else %}