OpenAPI: split import into URL and file commands with e2e tests (#6)

# TLDR;

Split the single OpenAPI import button into two: cloud icon for URL
import (downloads spec via HTTPS) and file icon for local file import.
Added context menu commands, response/request headers in panels, and e2e
tests with real network calls.

# Details

- **Two import commands**: `napper.importOpenApiUrl` (cloud icon)
downloads an OpenAPI spec from a URL; `napper.importOpenApiFile` (file
icon) picks a local spec file
- **URL download**: HTTPS with redirect following, error handling for
4xx responses
- **Context menus**: add-to-playlist, duplicate, delete, copy path,
enrich with AI
- **Response panel**: now shows request headers alongside response
headers
- **Playlist panel**: shows request/response headers per step
- **Report generator**: includes request/response headers in HTML
reports
- **F# Output.fs**: additional output formatting
- **E2e tests**: real network calls to petstore swagger URL, 404 error
handling, redirect following, temp file creation

# How do the tests prove the change works

- `downloadSpec fetches valid OpenAPI from petstore URL` — makes a real
HTTPS call to petstore3.swagger.io, parses the JSON, asserts it has
`openapi` version field and `paths` with entries
- `downloadSpec returns error for 404 URL` — hits
httpbin.org/status/404, asserts the result is an error with the correct
prefix
- `downloadSpec follows redirects` — uses httpbin.org redirect-to
endpoint, verifies the final response is valid OpenAPI
- `saveTempSpec writes file and returns path` — writes to a real temp
dir, reads back, asserts content matches
- `saveTempSpec overwrites existing file` — writes twice, asserts second
content wins
- Command registration tests verify both `napper.importOpenApiUrl` and
`napper.importOpenApiFile` are registered

---------

Author: MelbourneDeveloper

.github/workflows/pr.yml

@@ -31,12 +31,19 @@ jobs:
       - name: Build CLI, compile extension & tests
         run: npm run pretest
 
-      - name: Unit tests
+      - name: Unit tests with coverage
         run: npm run test:unit
 
       - name: E2E tests
         run: xvfb-run --auto-servernum npm test
 
+      - name: Upload TypeScript coverage
+        if: always()
+        uses: actions/upload-artifact@v4
+        with:
+          name: typescript-coverage
+          path: coverage/typescript/report/
+
   test-fsharp:
     name: F# Build & Tests
     runs-on: ubuntu-latest
@@ -47,14 +54,24 @@ jobs:
         with:
           dotnet-version: "10.0.x"
 
+      - name: Install ReportGenerator
+        run: dotnet tool install --global dotnet-reportgenerator-globaltool
+
       - name: Install dotnet-script
         run: dotnet tool install -g dotnet-script
 
       - name: Build
         run: dotnet build --nologo
 
-      - name: Test
-        run: dotnet test --nologo --verbosity normal
+      - name: Test with coverage
+        run: bash scripts/test-fsharp.sh
+
+      - name: Upload F# coverage
+        if: always()
+        uses: actions/upload-artifact@v4
+        with:
+          name: fsharp-coverage
+          path: coverage/fsharp/report/
 
   build-website:
     name: Website Build

.gitignore

@@ -1,29 +1,38 @@
-.too_many_cooks/
-.commandtree/
+# OS
+.DS_Store
 
+# .NET build output
 bin/
 obj/
 publish/
 
+# Node / TypeScript build output
+node_modules/
+dist/
+out/
+
 # VSCode extension
 src/Nap.VsCode/node_modules/
 src/Nap.VsCode/dist/
+src/Nap.VsCode/out/
 src/Nap.VsCode/*.vsix
+src/Nap.VsCode/.vscode-test/
 
-# Environment secrets
-.napenv.local
-
-# VSCode workspace settings
+# IDE settings
 .vscode/
 
-node_modules/
-dist/
-src/Nap.VsCode/.vscode-test/
-out/
-src/Nap.VsCode/out/
+# Secrets
+.napenv.local
 
-.DS_Store
+# Test output
+coverage/
+TestResults/
 
+# Tools
+.too_many_cooks/
+.commandtree/
 .playwright-mcp/
 
+# Generated files
 website/_site/
+examples/httpbin/advanced-report.html

Claude.md

@@ -28,6 +28,8 @@ You are working with many other agents. Make sure there is effective cooperation
 
 ### F# 
 - **Idiomatic F#**
+- **Standard F# result types** - Use the standard F# built-in result types
+- **Turn on F# analyzers** - Strict rules to enforce F# best practice
 
 ## Testing
 

examples/petstore/openapi.json

@@ -1,5 +1,22 @@
 #!/usr/bin/env bash
 set -euo pipefail
 cd "$(dirname "$0")/.."
-echo "==> Running F# tests..."
-dotnet test tests/Nap.Core.Tests --nologo
+
+COVERAGE_DIR="coverage/fsharp"
+rm -rf "$COVERAGE_DIR"
+mkdir -p "$COVERAGE_DIR"
+
+echo "==> Running F# tests with coverage..."
+dotnet test tests/Nap.Core.Tests --nologo \
+  --settings tests/Nap.Core.Tests/coverage.runsettings \
+  --results-directory "$COVERAGE_DIR/raw"
+
+echo "==> Generating coverage report..."
+reportgenerator \
+  -reports:"$COVERAGE_DIR/raw/*/coverage.cobertura.xml" \
+  -targetdir:"$COVERAGE_DIR/report" \
+  -reporttypes:"Html;TextSummary;Cobertura;lcov"
+
+echo ""
+echo "=== F# Coverage Summary ==="
+cat "$COVERAGE_DIR/report/Summary.txt"

scripts/test-fsharp.sh

@@ -4,10 +4,24 @@ cd "$(dirname "$0")/.."
 
 bash scripts/build-cli.sh
 
-echo "==> Compiling test TypeScript..."
+COVERAGE_DIR="coverage/typescript"
+rm -rf "$COVERAGE_DIR"
+mkdir -p "$COVERAGE_DIR"
+
+echo "==> Compiling TypeScript..."
 cd src/Nap.VsCode
 npm run compile
 npm run compile:tests
 
+echo "==> Running unit tests with coverage..."
+npx c8 \
+  --temp-directory "../../$COVERAGE_DIR/tmp" \
+  --report-dir "../../$COVERAGE_DIR/report" \
+  mocha out/test/unit/**/*.test.js --ui tdd --timeout 5000
+
 echo "==> Running e2e tests..."
 npx vscode-test
+
+echo ""
+echo "=== TypeScript Coverage Summary ==="
+echo "Report: $COVERAGE_DIR/report/index.html"

scripts/test.sh

@@ -667,55 +667,6 @@ nap/
 
 ---
 
-## OpenAPI / Swagger Test Generation (Phase 6+)
+## OpenAPI / Swagger Test Generation
 
-Nap will support generating `.nap` files and `.naplist` playlists from OpenAPI 3.x specs.
-
-```sh
-# Generate a collection from an OpenAPI spec
-nap generate openapi ./petstore.yaml --output ./petstore/
-
-# Generate from a URL
-nap generate openapi https://api.example.com/openapi.json --output ./generated/
-
-# Generate only for specific tags or paths
-nap generate openapi ./petstore.yaml --tag users --output ./users/
-```
-
-### What gets generated
-
-For each operation in the spec:
-- A `.nap` file with the correct method, URL template, and required headers.
-- Example request bodies derived from the spec's `example` / `examples` fields or schema defaults.
-- Basic `[assert]` blocks asserting the documented success status code and response shape.
-- Placeholder `{{variables}}` for path params, query params, and auth headers.
-
-```nap
-# Generated from GET /users/{userId}
-[meta]
-name        = "Get user by ID"
-description = "Auto-generated from petstore.yaml — operation: getUserById"
-tags        = ["users", "generated"]
-
-[vars]
-userId = "REPLACE_ME"
-
-[request]
-method = GET
-url    = {{baseUrl}}/users/{{userId}}
-
-[request.headers]
-Authorization = Bearer {{token}}
-
-[assert]
-status = 200
-body.id exists
-body.name exists
-```
-
-### Design notes
-
-- Generated files are **starting points**, not final tests. The user edits and extends them.
-- Re-running `nap generate` over an existing collection offers a `--diff` mode that shows what changed in the spec without overwriting custom assertions.
-- A `[meta] generated = true` flag marks auto-generated files so tooling can identify them.
-- Generation is a separate `Nap.OpenApi` library/package, keeping the core runtime free of OpenAPI dependencies.
+See [OpenAPI Generation Specification](./OpenApiGeneration.md) for the full specification covering one-click test suite generation from OpenAPI specs, including AI-assisted enrichment via GitHub Copilot.