generate_test_datasets.py

import json
import random
from datetime import datetime, timedelta

# ── PROFILES ──────────────────────────────────────────────────────────────────
# Each profile defines how a different team behaves

PROFILES = {
    "team_b_high_performing": {
        "team_name":    "Engineering Team B",
        "date_range":   "January 2024 – March 2024",
        "delivery_rate": 0.88,          # high delivery
        "large_sp_delivery": 0.78,      # even large items get done
        "state_multiplier": 0.65,       # faster than baseline
        "anomaly_sprints_qa": [],        # no QA anomalies
        "anomaly_sprints_cr": [],        # no code review anomalies
        "anomaly_assignee": None,
        "anomaly_assignee_sprints": [],
    },
    "team_c_struggling": {
        "team_name":    "Engineering Team C",
        "date_range":   "January 2024 – March 2024",
        "delivery_rate": 0.50,          # low delivery
        "large_sp_delivery": 0.25,      # large items rarely done
        "state_multiplier": 1.6,        # much slower than baseline
        "anomaly_sprints_qa": [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11],  # QA slow almost every sprint
        "anomaly_sprints_cr": [2, 4, 6, 8, 10],
        "anomaly_assignee": "alice",
        "anomaly_assignee_sprints": [3, 4, 5, 6, 7, 8, 9, 10],
    },
}

NUM_SPRINTS       = 12
ITEMS_PER_SPRINT  = (20, 30)
SPRINT_DURATION   = 14
START_DATE        = datetime(2024, 1, 1)

ISSUE_TYPES        = ["story", "bug", "task"]
ISSUE_TYPE_WEIGHTS = [0.45, 0.35, 0.20]
STORY_POINTS       = [1, 2, 3, 5, 8, 13]
STORY_POINT_WEIGHTS= [0.10, 0.20, 0.30, 0.25, 0.10, 0.05]
ASSIGNEES          = ["alice", "bob", "carol", "david", "eve"]

STATE_DURATIONS = {
    "story": {
        "todo":        (0.5, 2.0),
        "in_progress": (2.0, 6.0),
        "code_review": (1.0, 4.0),
        "qa":          (1.5, 5.0),
        "deployment":  (0.5, 1.5),
    },
    "bug": {
        "todo":        (0.5, 1.5),
        "in_progress": (1.0, 4.0),
        "code_review": (0.5, 3.0),
        "qa":          (1.0, 3.5),
        "deployment":  (0.5, 1.0),
    },
    "task": {
        "todo":        (0.5, 1.5),
        "in_progress": (1.0, 3.0),
        "code_review": (0.5, 2.0),
        "qa":          (0.5, 2.0),
        "deployment":  (0.5, 1.0),
    },
}

TITLES = {
    "story": [
        "Add user authentication flow", "Build dashboard overview page",
        "Implement export to CSV feature", "Create onboarding wizard",
        "Add team settings panel", "Integrate Slack notifications",
        "Build sprint summary view", "Add role-based permissions",
        "Implement search functionality", "Create API rate limiting",
    ],
    "bug": [
        "Fix login timeout issue", "Resolve broken pagination",
        "Fix date formatting on reports", "Resolve null pointer on assignee field",
        "Fix duplicate notifications bug", "Resolve sprint dates overlap",
        "Fix broken CSV export", "Resolve incorrect story point totals",
        "Fix missing state transitions", "Resolve timezone mismatch",
    ],
    "task": [
        "Update dependencies to latest versions", "Write unit tests for baseline engine",
        "Set up staging environment", "Document API endpoints",
        "Clean up legacy migration scripts", "Review and update error logging",
        "Configure CI/CD pipeline", "Audit database indexes",
    ],
}


def build_states(issue_type, sprint_start, sprint_idx, assignee, delivered, profile):
    states    = []
    curr_time = sprint_start + timedelta(hours=random.randint(0, 48))
    mult      = profile["state_multiplier"]

    for state in ["todo", "in_progress", "code_review", "qa", "deployment"]:
        min_d, max_d = STATE_DURATIONS[issue_type][state]
        min_d *= mult
        max_d *= mult

        if state == "qa" and sprint_idx in profile["anomaly_sprints_qa"]:
            max_d *= 2.2
        if state == "code_review" and sprint_idx in profile["anomaly_sprints_cr"]:
            max_d *= 2.0
        if assignee == profile.get("anomaly_assignee") and \
           sprint_idx in profile["anomaly_assignee_sprints"]:
            max_d *= 1.8

        duration = timedelta(days=random.uniform(min_d, max_d))
        entered  = curr_time
        exited   = curr_time + duration

        if not delivered and state == "qa":
            states.append({"state": state, "entered": entered.strftime("%Y-%m-%d"), "exited": None})
            break

        states.append({
            "state":   state,
            "entered": entered.strftime("%Y-%m-%d"),
            "exited":  exited.strftime("%Y-%m-%d"),
        })
        curr_time = exited

    if delivered:
        states.append({"state": "done", "entered": curr_time.strftime("%Y-%m-%d"), "exited": None})

    return states


def generate_dataset(profile_key):
    profile      = PROFILES[profile_key]
    all_issues   = []
    issue_counter = 100

    for sprint_idx in range(NUM_SPRINTS):
        sprint_start = START_DATE + timedelta(days=sprint_idx * SPRINT_DURATION)
        sprint_end   = sprint_start + timedelta(days=SPRINT_DURATION)
        sprint_id    = f"sprint_{sprint_idx + 1:02d}"
        num_items    = random.randint(*ITEMS_PER_SPRINT)

        for _ in range(num_items):
            issue_type   = random.choices(ISSUE_TYPES, weights=ISSUE_TYPE_WEIGHTS)[0]
            story_points = random.choices(STORY_POINTS, weights=STORY_POINT_WEIGHTS)[0]
            assignee     = random.choice(ASSIGNEES)

            base_rate = profile["delivery_rate"]
            if story_points >= 8:
                base_rate = profile["large_sp_delivery"]
            elif story_points >= 5:
                base_rate = (profile["delivery_rate"] + profile["large_sp_delivery"]) / 2

            # Struggling team gets progressively worse over time
            if profile_key == "team_c_struggling" and sprint_idx >= 6:
                base_rate *= 0.85

            delivered = random.random() < base_rate

            states = build_states(
                issue_type, sprint_start, sprint_idx,
                assignee, delivered, profile
            )

            all_issues.append({
                "issue_id":     f"PROJ-{issue_counter:03d}",
                "title":        random.choice(TITLES[issue_type]),
                "type":         issue_type,
                "story_points": story_points,
                "assignee":     assignee,
                "sprint_id":    sprint_id,
                "sprint_start": sprint_start.strftime("%Y-%m-%d"),
                "sprint_end":   sprint_end.strftime("%Y-%m-%d"),
                "committed":    True,
                "delivered":    delivered,
                "states":       states,
            })
            issue_counter += 1

    return all_issues, profile


if __name__ == "__main__":
    for profile_key in PROFILES:
        print(f"\nGenerating {profile_key}...")
        data, profile = generate_dataset(profile_key)

        output_path = f"data/{profile_key}.json"
        with open(output_path, "w") as f:
            json.dump(data, f, indent=2)

        delivered = sum(1 for i in data if i["delivered"])
        print(f"  {len(data)} issues, {delivered/len(data)*100:.1f}% delivery rate")
        print(f"  Saved to {output_path}")