What problem does it solve? Montréal's open data portal publishes dataset titles, descriptions, tags, and field names almost entirely in French, so English-language queries fail to find relevant datasets and results confuse English-speaking users. This Skill provides the translation patterns, field mappings, and formatting conventions needed to search in French and present results in the user's language. ## Core Features & Use Cases - Search Term Translation: Reference tables mapping English terms (trees, permits, transit, budget) to the French keywords that actually match CKAN catalog entries. - Bilingual Field Recognition: Patterns for detecting paired FR/EN columns (e.g., Essence_fr/Essence_ang, NOM_FR/NOM_AN) and handling French-only fields like ARROND_NOM. - Borough Name Reference: Complete mapping of the 19 borough codes (AHU, CDN, PLA...) to full names, with guidance on accent and dash-variant matching. - Use Case: A user asks "How many trees are in the Plateau?" The agent translates "trees" to "arbres", searches the catalog in French, filters records by ARROND_NOM = 'Le Plateau-Mont-Royal', and presents the answer in English. ## Quick Start Ask a question about Montréal open data in English, such as how many trees are in the Plateau, and have the agent search the French catalog and reply in English.
