Skip to content

validation.md

Latest commit

 

History

History
88 lines (66 loc) · 3.78 KB

validation.md

File metadata and controls

88 lines (66 loc) · 3.78 KB
Raw

✅ Data Alchemist – Validation Rules

This document outlines all the required and optional validation checks for the Data Alchemist application. These validations ensure the integrity, consistency, and usability of the uploaded datasets: clients.csv, workers.csv, and tasks.csv.

📁 File-wise Required Columns

clients.csv

  • ClientID
  • ClientName
  • PriorityLevel (1–5)
  • RequestedTaskIDs (comma-separated TaskIDs)
  • GroupTag
  • AttributesJSON (must be valid JSON)

workers.csv

  • WorkerID
  • WorkerName
  • Skills (comma-separated)
  • AvailableSlots (array of phase numbers)
  • MaxLoadPerPhase (integer)
  • WorkerGroup
  • QualificationLevel

tasks.csv

  • TaskID
  • TaskName
  • Category
  • Duration (≥ 1)
  • RequiredSkills (comma-separated)
  • PreferredPhases (list or range)
  • MaxConcurrent (integer)

🔍 Core Validation Rules

ID Validation Description
V1 Missing required column(s) in any uploaded file
V2 Duplicate IDs (ClientID, WorkerID, TaskID)
V3 Malformed lists (e.g., AvailableSlots not array of valid numbers)
V4 Out-of-range values – PriorityLevel not in 1–5, Duration < 1
V5 Broken JSON in AttributesJSON
V6 Unknown references – RequestedTaskIDs point to missing TaskIDs
V7 Circular co-run groups (e.g., A → B → C → A)
V8 Conflicting rules vs. phase-window constraints
V9 Overloaded workers – AvailableSlots.length < MaxLoadPerPhase
V10 Phase-slot saturation – sum of task durations in a phase > worker slots
V11 Skill-coverage matrix – each RequiredSkill must map to ≥1 worker
V12 Max-concurrency feasibility – MaxConcurrent exceeds available skilled workers

🤖 Optional AI-Powered Validations

ID Feature
A1 AI-assisted validation – catch data that "looks wrong" using heuristics
A2 Natural language data retrieval (e.g., "Tasks in phase 3 longer than 2 phases")
A3 Natural language rule creation (e.g., "T1 and T2 must run together")
A4 Rule recommendations based on patterns in uploaded data
A5 Data correction suggestions for broken JSON, malformed formats, etc.

🔁 Data Relationship Validations

Source Target Rule
Clients Tasks All RequestedTaskIDs must exist in tasks.csv
Tasks Workers All RequiredSkills must be covered by at least one worker
Workers Phases AvailableSlots must be valid and mapped
GroupTags Groups GroupTag (clients) and WorkerGroup must be consistent if used in rules

📌 Note

  • You must validate on initial upload and inline edits.
  • Highlight errors in the grid and show a validation summary panel.
  • You can extend these validations to make the app smarter using AI and natural language support.