Analyze RFE

Analyze RFE

Implementation guidance for /oape:analyze-rfe. When invoked, execute the steps below.

Prerequisites

• Jira: JIRA_PERSONAL_TOKEN set (and optionally JIRA_URL, default https://issues.redhat.com). If unset, prompt the user and exit.
• Input: RFE key (e.g. RFE-7841) or Jira URL. Extract key from URL if needed (e.g. https://issues.redhat.com/browse/RFE-7841 → RFE-7841).
• Optional (for Step 3 script-based context): GitHub CLI gh authenticated (gh auth login). Used by gather_component_context.py (which calls github_pr_analyzer.py and related scripts).

Step 1: Fetch the RFE

1. Parse input: if it looks like a URL, extract the issue key; otherwise use as key.
2. Fetch the RFE:
   • Preferred: Run the script (from plugin root or skill dir):

     python3 plugins/oape/skills/analyze-rfe/scripts/fetch_rfe.py {issue_key}
     

     Script requires JIRA_PERSONAL_TOKEN and pip install requests. It prints JSON to stdout and clear errors to stderr.
   • Alternative: Use curl or Python requests:

     curl -sS -H "Authorization: Bearer $JIRA_PERSONAL_TOKEN" \
          -H "Accept: application/json" \
          "$JIRA_URL/rest/api/2/issue/{key}?fields=summary,description,components,labels,status,issuetype"
     

3. On 401/403/404: report clear error and exit. If token missing, show setup instructions (see script stderr or command doc).
4. Extract: key, summary, description, components, status, labels.

Step 2: Parse RFE Content

From the description (strip Jira wiki markup), extract:

• Nature/Description — what is being requested
• Current Limitation — what doesn’t work today
• Desired Behavior — what should happen (bullets/paragraphs)
• Use Case — intended usage and scenarios
• Business Requirements — impact, justification
• Affected Components — teams, operators (from description and Jira components)

Note missing sections; still proceed with available content.

Step 3: Gather Component Context and Historic PR Analysis (Required)

1. Extract component names from RFE (from Jira components field and description).

2. Extract keywords from RFE description, summary, and desired behavior for PR search.

3. For each affected component, run the component context gatherer with PR analysis:

   python3 plugins/oape/skills/analyze-rfe/scripts/gather_component_context.py <component-name> \
     --keywords "keyword1" "keyword2" "keyword3" \
     --max-prs 50 \
     --deep-dive 3 \
     --analyze-upstream \
     --analyze-operands \
     -o .work/jira/analyze-rfe/<rfe-key>/component-<component-name>-context.md
   

   What this does:

   • Discovers downstream repo (e.g., openshift/cert-manager-operator)
   • Discovers upstream repo (e.g., cert-manager/cert-manager)
   • Searches for relevant PRs using RFE keywords (checks title, body, comments)
   • Analyzes PR history for design decisions, lessons learned, ADRs
   • If component is an operator, discovers and analyzes operand repos
   • Caches results to .work/jira/analyze-rfe/cache/ for performance

   Prerequisites: gh CLI authenticated (gh auth login). If gh is not available, log a warning and skip PR analysis (generate basic context only).

4. Fallback: Search workspace for **/context.md files if script-based analysis fails.

5. Synthesize: Merge PR insights, repo structure, and context.md data into the final report.

6. Use in output: Include "Component Context & Historical PR Analysis" section with:

   • What the component does, architecture pattern
   • Key implementation patterns from historic PRs
   • Relevant PRs with design rationale
   • Upstream vs downstream differences (if upstream analyzed)
   • Operand repos and their purpose (if operator)
   • Risk factors and recommended approach

Step 4: Generate EPIC(s)

• Count: One epic for a single capability; 2–3 if clearly distinct (e.g. API + UI, or phased MVP then enhancements).
• Per epic: Summary/title, objective, scope (in/out), 3–6 acceptance criteria (outcome-focused), target users.
• Align with the RFE’s desired behavior and use case.

Step 5: Generate User Stories

• Format: “As a <role>, I want to <action>, so that <value>.”
• Per story: User story text, short summary (5–10 words), 2–6 acceptance criteria (testable), outcome (value delivered).
• One story per distinct user-facing capability; right-size for one sprint.
• Map to desired behavior and (if present) workspace context key areas.

Step 6: Define Outcomes

• For each story: 1–2 sentence outcome (business/value delivered).
• Add epic-level outcome summary.

Step 7: Output the Report

1. Emit markdown in this structure:

# RFE Analysis: [KEY] - [Title]

## RFE Summary
| Field | Value |
|-------|-------|
| **Source** | [link] |
| **Key Capability** | ... |
| **Business Driver** | ... |
| **Affected Components** | ... |

## Component Context & Historic PR Analysis

[For each affected component]

### Component: [Component Name]

**Repositories**:
- Downstream: openshift/[repo-name]
- Upstream: [upstream-org/repo-name] (if applicable)
- Operands: [operand repos] (if operator)

**Architecture**: [Operator/CLI/Library/Service]

**Key Implementation Patterns** (from historic PRs):
1. [Pattern 1 from PR analysis]
2. [Pattern 2 from PR analysis]

**Relevant Historic PRs**:
- **PR #[number]** ([date]): [Title]
  - **Design Insight**: [What was designed and why]
  - **Scope**: [S/M/L - files changed]
  - **Relevance**: [Why this matters for current RFE]

**Upstream vs Downstream** (if upstream analyzed):
- [Key differences in architecture, features, or approach]

**Risk Factors**:
- [Risks identified from PR analysis and codebase structure]
- **Mitigation**: [Recommendations]

**Recommended Implementation Approach**:
- [Based on historic patterns and lessons learned]

## EPIC(s)
### EPIC 1: [Title]
**Objective**: ...
**Scope**: In scope / Out of scope
**Acceptance Criteria**: ...

## User Stories
### Epic 1 → Story 1.1: [Title]
**User Story**: As a ... I want ... So that ...
**Acceptance Criteria**: ...
**Outcome**: ...

## Outcomes Summary
| Story | Outcome |
...
---
*Generated by `/oape:analyze-rfe` on [timestamp]*

2. Optionally write the same content to .work/jira/analyze-rfe/<rfe-key>/breakdown.md (create directory if needed).

Error Handling

• No token: Explain how to create and set JIRA_PERSONAL_TOKEN; exit.
• Issue not found: Check key and permissions; exit.
• Not RFE project: Warn and continue.
• Sparse RFE: Note gaps; still produce best-effort breakdown.

Best Practices

• Synthesize from the RFE; don’t copy-paste long blocks.
• Epics = quarter-sized; stories = sprint-sized.
• Every story has a clear outcome.
• When workspace context exists, use it to tighten scope and key areas.