feat(showcase-ecommerce): add context docs for talk-to-data and operational scenarios#195
Merged
Merged
Conversation
|
The latest updates on your projects. Learn more about Vercel for GitHub. 1 Skipped Deployment
|
…tional scenarios
Adds 03-context.json — 18 Document entities organized in a browsable
folder hierarchy, linked to their source datasets via relatedAssets so
content is indexed alongside datasets for semantic search.
Document tree:
- E-Commerce Data Catalog (root)
- Core Transactional Tables (orders, order_items, customers, products,
inventories, promotions, addresses)
- Fulfillment & Reference Data (warehouses, product_categories, regions)
- Analytics Layer (order_details, order_history guides)
- Operational Runbooks (order count discrepancy, inventory staleness,
promotion attribution)
- Key Metrics Reference (GMV, AOV, CLV, return rate, fill rate)
Each document covers: purpose, column semantics, enum value definitions,
common query patterns, gotchas, and join maps — supporting talk-to-data,
operational debugging, and data development scenarios.
Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>
shirshanka
force-pushed
the
feat/showcase-ecommerce-context-docs
branch
from
e12d169 to
b319f79
Compare
nwadams
approved these changes
Apr 21, 2026
4 tasks
nwadams
pushed a commit
that referenced
this pull request
Jun 8, 2026
…chor demo (#198) Adds a corpus of urn:li:query MCPs to back the anchor-generation pipeline discussed in the talk-to-data demo work that #195 set up the context docs for. Until now the showcase-ecommerce datapack had no query entities, so the anchor pipeline had no SQL to fingerprint and rank. What's added: - 04-queries.json — 1,246 MCPs spanning 312 urn:li:query entities: * 160 "cluster" queries across 50 distinct analyst intents (3–5 variants per intent: raw vs. analytics.order_details mart, CTE vs. subquery, alternate join paths, a few intentional anti-shapes) * 152 "noise" queries — one-off analyst drill-downs against specific customer/order/product/promotion IDs, date-window investigations, DQ checks, schema exploration, abandoned half-finished queries * Each query carries queryProperties, querySubjects (auto-detected dataset URNs), dataPlatformInstance (snowflake/b2fd91), and subTypes * created/lastModified timestamps spread across the data window so the corpus looks like a real activity history * authors mapped to the 11 corpuser entities already in 02-data.json (Sarah Chen, Andrea Garcia, David Kim, Ben Porter, Julia Novak, James Wilson, Karen Okonkwo, Ian Chen) - queries/ — readable source SQL alongside the build artifact: * One directory per intent with a 00-intent.yaml (question, canonical definition, variant axes, data caveats) plus the .sql files * noise/ subdirectory with the 152 long-tail queries * intent-matrix.yaml — top-level index across all 50 intents * README.md — contributor guide including data-realities gotchas (e.g. orders.order_status is NUMBER 0..4 not strings, mailshot is 0/1 not Y/N, customers.country_id isn't a real FK) - index.json — adds 04-queries.json to the file list The corpus was authored against the real ORDER_ENTRY_DB on snowflake (every query was executed against the warehouse during authoring) and surfaces a few demo-grade data anomalies on the way: - 4028/5027 promoted orders fall outside their promo's date window - 1250 In-Store Pickup orders carry a delivery_address_id - 3 warehouses show NEGATIVE avg dispatch lead time - Some promotions have negative discount rates (unit_price > list_price) - Order 9866 header_total = $210 but its line items sum to $1000 Verified via `datahub datapack load --dry-run` — all MCPs schema-compatible. Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
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
2 participants
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
Adds
03-context.jsonto theshowcase-ecommercedatapack — 18 Document entities organized in a browsable folder hierarchy, each linked to their source datasets viarelatedAssetsso the content is indexed alongside datasets for semantic search.Document tree (54 MCPs):
Each document covers: table purpose, column semantics with enum value definitions, known gotchas, common query patterns, and join maps.
Designed for three scenarios:
Test plan
showInGlobalContext: trueon all 18 docs)🤖 Generated with Claude Code