odoo-indexer

name

Odoo Indexer

description

Fast indexer for Odoo codebases - 95% more token-efficient than reading files. USE AUTOMATICALLY AND PROACTIVELY before ANY Odoo code work. AUTO-TRIGGER when user mentions models (sale.order, res.partner, account.move, etc.), fields (partner_id, name, state, etc.), views (form, tree, kanban), XML IDs, or when you need to search/validate/explore Odoo code. USE BEFORE writing code to validate references exist, USE BEFORE reading files to locate elements, USE DURING debugging to trace dependencies. CRITICAL: Always validate models/fields/xmlids with indexer before using them in code. Provides <100ms searches vs reading 20+ files. Essential for all Odoo development tasks.

Odoo Indexer Skill

⚡ PROACTIVE USAGE: This skill should be used automatically when user asks questions about Odoo code structure, models, fields, views, or modules.

Fast indexing and search for Odoo codebase elements with sub-100ms query performance.

---

When to Auto-Use This Skill

✅ User asks: "What is sale.order?" → AUTO-USE: uv run scripts/get_details.py model "sale.order"

✅ User asks: "What fields does res.partner have?" → AUTO-USE: uv run scripts/get_details.py model "res.partner"

✅ User asks: "Find all Many2one fields in sale module" → AUTO-USE: uv run scripts/search_by_attr.py field --filters '{"field_type":"Many2one"}' --module sale

✅ User asks: "Where is project.task defined?" → AUTO-USE: uv run scripts/search.py "project.task" --type model

✅ User asks: "Does sale.order have partner_id field?" → AUTO-USE: uv run scripts/search.py "partner_id" --type field --parent "sale.order"

✅ User asks: "Search for task views" → AUTO-USE: uv run scripts/search.py "%task%" --type view

✅ User asks: "List all modules" → AUTO-USE: uv run scripts/list_modules.py

✅ User asks: "Tell me about the sale module" → AUTO-USE: uv run scripts/module_stats.py sale

---

Trigger Keywords

Automatically use this skill when user query contains:

• "what is" + model/field/view name
• "what fields" + model name
• "find" / "search" / "search for" + element type
• "show me" + element name
• "where is" + element name
• "does...have" + model + field
• "list" + element type
• "tell me about" + module name
• "how does" + model/method name
• "get" / "display" + element info

---

Prerequisites

This skill requires uv (Python package manager). Usually auto-installed by /odoo-setup.

Manual install (if needed):

# macOS/Linux
curl -LsSf https://astral.sh/uv/install.sh | sh

# Or via Homebrew (macOS)
brew install uv

---

Configuration

Environment variables (optional - auto-detected by skill):

export ODOO_PATH="$HOME/letzdoo-sh/odoo/custom/src"  # Auto-detected if not set
export SQLITE_DB_PATH="$HOME/.odoo-indexer/odoo_indexer.sqlite3"  # Default location

The skill auto-detects common Odoo locations:

1. $ODOO_PATH (if set)
2. /home/coder/letzdoo-sh/odoo/custom/src
3. $HOME/odoo/custom/src
4. ./odoo/custom/src

---

Initial Setup

Automatic (via /odoo-setup command):

• Index is built automatically during plugin setup
• No manual steps needed

Manual (if index not built):

cd odoo-doodba-dev/skills/odoo-indexer
uv run scripts/update_index.py --full

Expected output:

Starting full index rebuild...
Scanning modules... Found 156 modules
Indexing: odoo (core) ████████████ 100%
Indexing: sale ████████████ 100%
...
✓ Indexing complete!
Summary:
- Modules: 156
- Models: 1,247
- Fields: 15,832
Database size: 42.3 MB

Indexing time:

• Small (50 modules): 1-2 minutes
• Medium (100-150 modules): 2-5 minutes
• Large (200+ modules): 5-10 minutes

---

Core Commands

All commands use uv run from the skills/odoo-indexer/ directory:

1. Search for Elements

# Search models
uv run scripts/search.py "sale.order" --type model

# Search fields in a module
uv run scripts/search.py "partner_id" --type field --module sale

# Search with wildcards (supports SQL LIKE patterns)
uv run scripts/search.py "action_view_%" --type action

# Limit results for performance
uv run scripts/search.py "sale%" --type model --limit 10

Parameters:

• query: Search term (supports % wildcards)
• --type: Element type (model/field/function/view/action/menu)
• --module: Filter by module name
• --parent: Filter by parent (e.g., model name for fields)
• --limit: Max results (default: 20)
• --offset: Pagination offset (default: 0)
• --json: Output as JSON

Performance: <50ms for exact match, <100ms for wildcard search

---

2. Get Element Details

Get complete information about a specific element:

# Model details (all fields, methods, views, actions)
uv run scripts/get_details.py model "sale.order"

# Field details (type, attributes, usage)
uv run scripts/get_details.py field "partner_id" --parent "sale.order"

# View details (structure, inheritance)
uv run scripts/get_details.py view "sale_order_form_view" --module sale

Returns:

• Full element definition
• All attributes and properties
• Related elements (views, actions, menus)
• File location and line numbers
• Usage statistics

Performance: <30ms

---

3. List Modules

# List all indexed modules
uv run scripts/list_modules.py

# Filter by pattern
uv run scripts/list_modules.py --pattern "sale*"

Returns:

• Module name
• Number of models, fields, views
• Installation status
• Location

---

4. Module Statistics

# Detailed module info
uv run scripts/module_stats.py sale

Returns:

• Total elements (models, fields, views, actions, menus)
• Main models with field counts
• Dependencies and dependents
• Installation path
• Index status

---

5. Find References

Find all locations where an element is used:

# Find all references to a model
uv run scripts/find_refs.py model "sale.order"

# Find field references
uv run scripts/find_refs.py field "partner_id" --ref-type definition

Reference types:

• definition - Where element is defined
• inheritance - Where element is inherited
• override - Where element is overridden
• reference - Where element is referenced

---

6. Search by Attribute

Advanced search using element attributes:

# Find all Many2one fields
uv run scripts/search_by_attr.py field --filters '{"field_type": "Many2one"}'

# Find transient models
uv run scripts/search_by_attr.py model --filters '{"model_type": "transient"}' --module sale

# Find form views
uv run scripts/search_by_attr.py view --filters '{"view_type": "form"}'

# Combine filters
uv run scripts/search_by_attr.py field --filters '{"field_type": "Many2one", "required": true}' --module sale

Common filters:

• Fields: field_type, required, readonly, compute, store
• Models: model_type (model/transient/abstract), inherit
• Views: view_type (form/list/search/kanban/calendar/pivot/graph)

---

7. Search XML IDs

# Search XML IDs with wildcards
uv run scripts/search_xml_id.py "action_view_%"

# Exact XML ID in module
uv run scripts/search_xml_id.py "sale_order_form_view" --module sale

Returns:

• Full XML ID (with module prefix)
• Element type (action/view/menu/etc)
• Target model (if applicable)
• File location

Important: Always use the full XML ID returned (with module prefix) in ref() calls.

---

8. Update Index

Keep the index up-to-date with code changes:

# Incremental update (fast - only changed files)
uv run scripts/update_index.py

# Full rebuild (slow - all files)
uv run scripts/update_index.py --full

# Index specific modules
uv run scripts/update_index.py --modules "sale,account,stock"

# Clear and rebuild
uv run scripts/update_index.py --clear --full

When to update:

• After installing new modules
• After major code changes
• After git pull with many changes
• Weekly (for active development)

Performance:

• Incremental: <30 seconds
• Full rebuild: 2-10 minutes (depending on size)

---

9. Index Status

Check index health and statistics:

uv run scripts/index_status.py

Returns:

• Database size
• Total elements indexed
• Last update timestamp
• Index health metrics
• Stale files (if any)

---

Performance Metrics

Operation	Time	vs File Reading
Exact model search	<20ms	50x faster
Field lookup	<30ms	30x faster
Wildcard search	<100ms	20x faster
Module stats	<50ms	40x faster
Reference search	<150ms	15x faster

Token efficiency: 95% fewer tokens than reading files

---

Usage Tips

1. Use Wildcards for Broad Searches

# Good - specific
uv run scripts/search.py "sale.order" --type model

# Also good - controlled wildcard
uv run scripts/search.py "sale.%" --type model --limit 10

# Avoid - too broad (slow)
uv run scripts/search.py "%" --type model  # Returns thousands of results

2. Limit Results for Performance

Always use --limit for wildcard searches:

uv run scripts/search.py "%partner%" --type field --limit 20

3. Combine Filters

Use specific filters to narrow results:

uv run scripts/search.py "partner_id" --type field --parent "sale.order" --limit 1

4. Cache Common Queries

For frequently accessed info, mention you're using cached data:

I've already looked up sale.order (from indexer cache).

5. Prefer Exact Searches

Exact searches are much faster:

# Fast
uv run scripts/get_details.py model "sale.order"

# Slower
uv run scripts/search.py "%sale.order%" --type model

---

Examples

Example 1: Model Overview

Query: "What is sale.order?"

Command:

uv run scripts/get_details.py model "sale.order"

Response format:

• Model name and description
• Module location
• Inheritance chain
• Key fields (10-15 most important)
• View count
• Action references
• File location

---

Example 2: Field Search

Query: "What fields does res.partner have?"

Command:

uv run scripts/get_details.py model "res.partner"

Response format:

• Field count
• Categorized fields (contact info, address, relationships)
• Field types and attributes
• Mention of full list availability

---

Example 3: Advanced Search

Query: "Find all Many2one fields pointing to res.partner"

Command:

uv run scripts/search_by_attr.py field --filters '{"field_type": "Many2one", "comodel_name": "res.partner"}' --limit 50

Response format:

• Grouped by module/model
• Field names with context
• Total count
• Query time

---

Troubleshooting

Index Not Found

Error: Database file not found

Solution:

cd odoo-doodba-dev/skills/odoo-indexer
uv run scripts/update_index.py --full

Or run /odoo-setup to rebuild everything.

---

Wrong Odoo Path

Error: Odoo directory not found

Solution:

# Set correct path
export ODOO_PATH="/path/to/your/odoo/custom/src"

# Verify
ls $ODOO_PATH/odoo  # Should show Odoo core files

# Rebuild index
uv run scripts/update_index.py --full

---

Stale Results

Issue: Search returns outdated information

Solution:

# Quick update (incremental)
uv run scripts/update_index.py

# Full rebuild if issues persist
uv run scripts/update_index.py --full

---

Slow Queries

Issue: Searches taking >1 second

Causes:

1. Too broad search pattern (use --limit)
2. Database corruption (rebuild with --full)
3. Large database (optimize queries)

Solution:

# Add limit
uv run scripts/search.py "%" --type model --limit 10

# Or rebuild index
uv run scripts/update_index.py --clear --full

---

Missing Dependencies

Error: ModuleNotFoundError: lxml

Solution:

cd odoo-doodba-dev/skills/odoo-indexer
uv sync  # Auto-installs from pyproject.toml

---

Integration with Commands

Used by /odoo-search

The /odoo-search command is a user-friendly wrapper around this skill:

• Interprets natural language queries
• Translates to appropriate indexer commands
• Formats results for readability

Used by Agents

odoo-developer agent:

• Validates all model/field references
• Researches codebase before implementation
• Checks XML IDs before use

odoo-verifier agent:

• Validates module structure
• Checks security groups
• Verifies view references

---

Best Practices

1. Always Validate References

Before using any model/field/XML ID in code:

# Verify model exists
uv run scripts/search.py "custom.model" --type model

# Verify field exists
uv run scripts/search.py "custom_field" --type field --parent "custom.model"

# Verify XML ID exists
uv run scripts/search_xml_id.py "custom_action" --module custom

2. Use Exact Names

Always use exact names from indexer (don't guess):

# Get exact field name
uv run scripts/get_details.py field "partner" --parent "sale.order"
# Returns: partner_id (not "partner")

3. Check Module Prefix

Always use module prefix from indexer for XML IDs:

uv run scripts/search_xml_id.py "action_view_task"
# Returns: project.action_view_task
# Use: ref('project.action_view_task')  # Correct
# NOT: ref('action_view_task')  # Wrong

4. Batch Validations

For multiple elements, batch searches:

# Instead of multiple calls, use one search
uv run scripts/search.py "partner_id|user_id|team_id" --type field --parent "sale.order"

5. Keep Index Fresh

Update index regularly:

# Daily incremental update
uv run scripts/update_index.py

# Weekly full update
uv run scripts/update_index.py --full

---

Critical Rules for Agents

When agents use this skill:

1. ALWAYS validate references before code generation
2. NEVER assume field names - get exact names from indexer
3. NEVER guess XML ID module prefixes - use indexer results
4. Use get_details.py for comprehensive info
5. Use search.py for finding elements
6. Use search_by_attr.py for filtered searches
7. Report query time in responses
8. Indexer is 95% more efficient than reading files

---

Database Details

Location: ~/.odoo-indexer/odoo_indexer.sqlite3 (default)

Size: 10-50 MB (varies with codebase size)

Schema:

• modules - Module metadata
• models - Model definitions
• fields - Field definitions
• views - View definitions
• actions - Action definitions
• menus - Menu structure
• xml_ids - XML ID mappings
• references - Cross-references

Indexed data:

• Python files (AST parsing)
• XML files (lxml parsing)
• CSV files (security rules)

---

Remember: This skill should be your FIRST choice for any Odoo code exploration. It's 95% faster and more accurate than reading files!