- Rename project folder Apis/email-api-data → Apis/email-data
- Rename csproj file: email-api-data.csproj → email-data.csproj
- Update csproj properties: AssemblyName and RootNamespace (email-data, Email.Data)
- Update C# namespaces: EmailApi.Data → Email.Data across all email-data files
- Update project references in api.csproj and email-api.csproj
- Update migration assembly references in api/Program.cs and email-api/Program.cs
- Update cv-search-job references to use email-data project and Email.Data namespace
- Update solution file to reference new email-data project path
- Remove hardcoded schema name from SmtpEmailDispatcher, use template service instead
This maintains consistency with other data project naming convention (no service-type suffix).
All tests passing, build succeeds.
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
- Delete cv-search-models/Data (duplicate of cv-search-data/Data)
- Delete myai-models/Data (duplicate of myai-data/Data)
- DbContext and Entities belong only in -data projects, not -models
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
- Update MigrationConstants.SchemaName in email-api-data from 'emailApi' to 'email'
- All migrations automatically use the new schema name via MigrationConstants reference
- Aligns with naming convention: 'email', 'rag', 'cvMatcher', 'cvSearch', 'myAi'
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
- Move IRagRepository, EfRagRepository, and VectorSerializer from rag-api/Data to rag-data/Repositories
- Add rag-api-models ProjectReference to rag-data.csproj for model type availability
- Delete rag-api/Data folder (no longer needed; all data access is now in rag-data)
- This aligns RAG with email-api and other services: all data code in the data project
Pattern: rag-api (API logic) → rag-data (repository, EF entities, migrations)
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
Added inline comments throughout myai.css to:
- Clarify complex CSS selectors (input:not selectors, specificity explanations)
- Explain design patterns (.is-invalid error state pattern)
- Document focus and error states
- Describe layout decisions (sticky result panel, hamburger dropdown)
- Clarify responsive breakpoints and what changes at each
- Explain the relationship between CSS and JS (e.g., .is-open, .is-invalid)
Comments are strategic and concise—added to complex/non-obvious sections
without bloating the file. All CSS rules remain unchanged—purely additive
documentation.
Key sections now have better context:
- Form field selectors and state handling
- Hamburger menu responsive behavior
- Result panel sticky positioning strategy
- Responsive grid layout changes at 900px and 560px
- Cookie banner and loader overlay behavior
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
Refactored legal.js from 135 → 124 lines (8% reduction) by:
- Removing local browserLang() and getLang() that are now in utils
- Simplifying to focus on page-specific injection logic
Kept legal page-specific functionality:
- Local LANG_KEY storage for page language preference
- injectTopbar() with language switcher buttons
- injectFooter() with language-aware copyright and legal links
- Event delegation for language link clicks
- DOMContentLoaded handler
Added clear JSDoc comments explaining the injection pattern and
how legal pages dynamically reuse common UI elements while supporting
language switching via event delegation.
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
Refactored cv-matcher.js from 257 → 213 lines (17% reduction) by:
- Removing duplicate helper functions now in utils/form-helpers.js
- Removing duplicate i18n logic now in utils/i18n.js
- Removing API loading code now in utils/api.js
Kept CV matcher-specific logic:
- CV file input change handler
- Async CV upload and match flow with two captcha tokens
- Match result rendering with score badge and lists
- escapeHtml() XSS prevention utility
- $(window).on('load') to load reCaptcha
All function calls updated to use window.MyAi.* utilities for consistency.
Added detailed JSDoc comments explaining the two-step async flow.
Updated cv-matcher/index.html to load all utilities before cv-matcher.js.
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
Refactored main.js from 544 → 266 lines (51% reduction) by:
- Removing duplicate functions now in utils/form-helpers.js
- Removing duplicate i18n logic now in utils/i18n.js
- Removing API loading code now in utils/api.js
- Removing cookie consent handlers now in modules/cookie-consent.js
Kept only page-specific form handlers:
- Contact form submission with reCaptcha
- Subscribe form submission with reCaptcha
- Language switcher initialization
- Footer year and version display
All calls now use window.MyAi.* utilities for consistency.
Updated index.html to load all utilities before main.js.
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
Create reusable utility modules to eliminate duplication across main.js,
cv-matcher.js, and legal.js:
- js/utils/form-helpers.js: showFieldError, clearFieldErrors, isValidEmail,
extractApiError — shared form validation and error handling
- js/utils/i18n.js: currentLang, t, applyLanguage, updateLegalLinks,
browserLang — shared translation and language switching
- js/utils/api.js: checkApiLive, getRecaptchaWebKey, getGoogleTagManagerId,
loadGoogleTagManager — shared API configuration loading
- js/modules/cookie-consent.js: getConsent, setConsent, initConsent,
setupConsentHandlers — cookie banner and consent management
All utilities exposed on window.MyAi namespace for use by existing pages.
Full JSDoc headers and inline comments for maintainability.
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
Each DbContext now explicitly configures its migration history table to use
the schema-qualified name pattern [schemaName].[_Migrations]:
- [cvMatcher].[_Migrations] for CvMatcherDbContext
- [emailApi].[_Migrations] for EmailApiDbContext
- [cvSearch].[_Migrations] for CvSearchDbContext
- [rag].[_Migrations] for RagDbContext
- [myAi].[_Migrations] for MyAiDbContext
This is done via OnConfiguring() with UseSqlServer().MigrationsHistoryTable(name, schema).
Removed incorrect rename migrations that were created due to misunderstanding
of the proper EF Core configuration approach.
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
- Rename EmailApiDbContext MigrationTableName from '_EmailApiMigrations' to '_Migrations'
- Rename MyAiDbContext MigrationTableName from '_MyAiMigrations' to '_Migrations'
- Add migrations to rename tables in database: emailApi._EmailApiMigrations → emailApi._Migrations, myAi._MyAiMigrations → myAi._Migrations
- Aligns with naming convention used in other schemas (cvMatcher, cvSearch, rag)
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
Remove Seed() call from CreateEmailTemplates Up() method to prevent
duplicate key violation when applying SeedEmailTemplates migration.
The original migration was attempting to seed data during schema creation,
but data seeding is now handled by the separate SeedEmailTemplates migration
(20260528130652). Keeping both Seed() calls caused PRIMARY KEY violation on
(email.html-shell.start, *) when the second migration tried to insert
already-existing templates.
This maintains the migration order: schema creation first, then data seeding
in a separate, dedicated migration.
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
- Create SeedEmailTemplates migration (20260528130652) with all email templates
- Add Microsoft.EntityFrameworkCore.Design to email-api.csproj for EF migrations
- Add EmailApiDbContext registration and migration support to email-api Program.cs
- Configure IEmailTemplateRepository and IEmailTemplateService in email-api
- All 14 email templates now seeded in emailApi schema (HTML shells, CV match, job search)
- Templates include proper placeholder support ({{score}}, {{count}}, {{jobLabel}}, etc.)
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
Both api and cv-search-job need to connect to email-api for sending emails.
Add EmailApi section to their appsettings.json with BaseUrl and InternalApiKey
placeholders. Environment variables from docker-compose will populate these at runtime.
Also add EmailApi credentials to docker-compose/.env:
- EmailApi__BaseUrl=http://email-api:8080
- EmailApi__InternalApiKey=<shared key>
- EmailApi__RequireApiKey=true
This ensures both services can authenticate and call the email-api service.
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
Updated general-dev-workflow skill to stress that for Docker projects,
BOTH build success AND successful container startup are required before
considering a change complete.
Key additions:
- Docker container startup verification is mandatory, not optional
- Containers must have status 'Up', not 'Restarting' or 'Exited'
- Check container logs for errors during startup
- Missing config files, invalid env vars, DB issues only show at runtime
- Startup failures block production deployments
- Updated Phase 4 checkpoint to include container startup validation
Lesson from email-api issue: missing appsettings.json caused restart loop
that had no visible logs during build. This must be caught in Phase 4
before code review and prevents production issues.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Merges refactor/web-cleanup-31 into main with all infrastructure fixes:
- Directory.Packages.props handling in all Dockerfiles
- Missing email-api-data dependencies in build configs
- Missing appsettings.json for email-api
- myai-smoke-test skill for automated end-to-end testing
- general-dev-workflow skill updates for build verification
All 7 containers building and running successfully.
Closes#33
The email-api service was missing its configuration file, which is required
for Serilog logging setup, database connection, and SMTP settings.
Created appsettings.json with:
- Serilog configuration for console, file, and email logging
- Database connection settings
- SMTP configuration for email sending
- Internal API key configuration
- File storage path configuration
This fixes the container crash loop caused by missing configuration.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Updated Phase 4 (Test) to include mandatory build verification:
- dotnet build for .NET projects
- docker compose --build for Docker projects
- Catch missing dependencies and configuration issues early
- Prevent build failures during code review and CI/CD
Also added tip about verifying builds before opening PR, and updated
Phase 4 checkpoint to include successful build requirement.
Lesson learned from docker build issues: catching these early saves
reviewers and CI/CD time, and prevents 'works on my machine' problems.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The email-api service was missing from the CI/CD build pipeline. Added:
- EMAIL_API_IMAGE environment variable
- Build step for email-api Dockerfile
- Push step for email-api image to registry
This ensures email-api images are built and pushed alongside other services
during the staging build workflow.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The api.csproj references both email-api-data and email-api-models, but the
Dockerfile was not copying them. This caused compilation warnings and potential
build failures.
Added COPY commands for both projects before restore and publish steps.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The cv-search-job.csproj references both email-api-data and email-api-models,
but the Dockerfile was not copying them into the build context. This caused
compilation errors about missing EmailApi namespace types.
Added COPY commands for both projects before restore and publish steps.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The email-api.csproj references email-api-data as a project dependency,
but the Dockerfile was not copying it into the build context. This caused
'Skipping project' warnings during restore/publish.
Added COPY commands for both .csproj (before restore) and source directory
(before publish) to include email-api-data in the Docker build.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The explicitly added versions were conflicting with the centralized version
definitions in Directory.Packages.props. Removed all explicit versions from:
- web/web.csproj
- Jobs/cv-cleanup-job/cv-cleanup-job.csproj
- Jobs/cv-search-job/cv-search-job.csproj
NuGet will now resolve versions from Directory.Packages.props which has the
canonical version definitions for the entire solution.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The Docker builds were failing because the centralized package version
management file (Directory.Packages.props) was not being copied into the
build context. This file is required for NuGet to resolve package versions
in projects that don't specify explicit versions.
Updated all Dockerfiles to copy Directory.Packages.props before running
dotnet restore:
- Apis/api/Dockerfile
- Apis/cv-matcher-api/Dockerfile
- Apis/rag-api/Dockerfile
- Jobs/cv-cleanup-job/Dockerfile
- Jobs/cv-search-job/Dockerfile
- web/Dockerfile
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The PackageReference items were missing version specifications:
- Microsoft.VisualStudio.Azure.Containers.Tools.Targets (added 1.21.0)
- Yarp.ReverseProxy (added 2.2.0)
This fixes the NuGet error NU1015 that prevented Docker builds from
successfully restoring package dependencies.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The smoke test now correctly includes docker-compose.override.yml which
configures local image builds instead of pulling from remote registry.
This fixes the 'failed to resolve reference' error when building containers
locally.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Creates a new skill that automates smoke testing of the myAi CV Matcher:
- Starts Docker Compose and waits for app health check
- Uploads CV.pdf and job description via Selenium WebDriver
- Verifies CV analysis results display (score, strengths, gaps)
- Confirms match email was sent by checking container logs
- Returns pass/fail summary with any failures detailed
Includes SKILL.md documentation and run_smoke_test.py automation script
with hardcoded test data (CV file path, job description). Can be extended
to test against different CVs/job descriptions via environment variables.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
legal.js now:
- Dynamically injects a common topbar (logo + language switcher) on all 6 pages
- Dynamically injects a language-aware footer (EN vs RO copyright text)
- Detects page language and builds appropriate language links
- Uses event delegation for language links (works on injected elements)
- Persists language preference to localStorage
All 6 legal HTML pages now:
- Removed the hardcoded topbar div (12 lines of identical HTML per file)
- Removed the hardcoded footer div (7 lines of HTML with language-specific content)
- Total savings: 114 lines of duplicated HTML across 6 pages
- Pages are 38% smaller (60 lines → 37 lines core content)
Closes#31
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Extract CV matcher form logic from main.js into dedicated cv-matcher.js:
- CV file upload with validation
- Job URL/description input
- reCaptcha token retrieval for upload and match actions
- Match result rendering with score badge, strengths, gaps, evidence
- Helper functions: extractApiError, escapeHtml, showFieldError, etc.
Depends on jQuery, i18n.js, and shared utilities from main.js.
Closes#31
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Extract i18n dictionary from main.js into dedicated i18n.js module.
Sets window.MyAi.i18n with 228 keys across English and Romanian.
Main.js will consume this via t(key) helper function.
Closes#31
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Remove all 7 inline style="display:none;" attributes from loaders/cookie elements
(now handled by CSS with .loader-overlay { display: none } and
.loader-overlay.loader-visible { display: flex })
- Remove orphan footer-legal class (unused in CSS) from footer-links divs
- Add <script src="/js/i18n.js"></script> before main.js on both pages
- Add <script src="/js/cv-matcher.js"></script> after main.js on cv-matcher page
jQuery 4.0 now detects CSS display:none correctly via getComputedStyle,
so class-based visibility (.loader-visible) works seamlessly.
Closes#31
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add section block comments throughout for navigability
- Merge two duplicate @media (max-width:900px) blocks into one
- Remove dead .ai-mark rule (replaced by <img> logo, never rendered)
- Move .cookie-overlay, .cookie-manage to display:none by default
(removes need for inline style="display:none;" on those elements)
- Add .loader-overlay.loader-visible{display:flex} so JS can use
class toggling instead of .css('display','flex') — works correctly
with jQuery 4.0's getComputedStyle-based visibility detection
- Consolidate brand-mark, console-line, nav-actions rules next to
their related blocks (were scattered at end of file)
Closes#31
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Bootstrap CSS (232 KB) and Bootstrap JS (39 KB) were loaded for only
4 utility classes (.btn-sm, .btn-dark, .btn-warning, .shadow).
Replace with 4 targeted rules in myai.css and delete both files.
style.css (784 lines) was unreferenced by any HTML page — delete it.
Closes#31
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The vendor file was misnamed jquery-1.12.4.min.js but contained v3.6.1.
Replaced with the correctly-named jquery-4.0.0.min.js (78 KB, up from 89 KB).
Compatibility check: none of the removed jQuery 4.0 APIs are used in
main.js — no $.isArray, $.isFunction, $.trim, $.type, $.parseJSON, etc.
All ajax/deferred/DOM methods in use ($.ajax, $.when, .done/.fail/.always,
.on, .prop, .css, .addClass etc.) remain in the full 4.0 build.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- UseJsonExceptionHandler now maps InvalidOperationException to 400 (was 500),
so upstream business-rule rejections reach the browser as actionable messages.
- CvMatcherController forwards Refit 4xx bodies from cv-matcher-api instead
of swallowing them in a generic 502.
- ErrorResponse.Score removed; CaptchaController puts the score in Detail.
- Frontend extractApiError helper reads the server Error/error/title field for
4xx responses and falls back to a generic i18n string for 5xx / missing body.
- All four failure handlers (CV upload, CV match, contact form, subscribe form)
updated to use extractApiError with the correct rate-limit i18n key.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Update CLAUDE.md: replace incorrect 'no XML doc on internal code' rule
with the correct convention (XML doc on all public methods and
non-trivial private/protected helpers)
- Restore /// <summary> on FileDownloadController private helpers
(HandleRangeRequest, StreamRangeAsync)
- Add full XML doc to all service contracts:
ICaptchaVerifier, IEmailSender, ICvMatcherService, IJobTextExtractor,
IJobTokenService, IDocumentClassifier, IRagService, ITextChunker,
ITextExtractor, IEmailTemplateService, ITemplateService
- Add /// <summary> and /// <inheritdoc /> to all concrete service classes
and their methods: RecaptchaVerifier, EmailApiEmailSender,
SmtpEmailDispatcher, CvMatcherService, JobTextExtractor, JobTokenService,
RagService, DocumentClassifier, TextChunker, TextExtractor,
HtmlJobSearcher, CvSearchEmailSender, CvSearchJobTask,
EmailTemplateService, DbTemplateService
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- EmailController: add class summary, full SwaggerResponse/ProducesResponseType
for 400 and 500, and Description on SwaggerOperation
- ContactController: fix terse "Failed." error message to
"Could not process subscription."
- FileDownloadController: remove redundant XML <response code> tags from
the public action doc block; convert private-method /// <summary> to //
(project convention: no XML doc on internal code)
- CvMatcherService: remove two dead commented-out blocks (old email send
and BuildEmailBody helper)
- JobTokenService: comment the phone/contact-line regex filter in
ExtractKeywords
- DocumentClassifier: comment the keyword-frequency scoring approach and
the confidence formula
- TextChunker: comment the sliding-window step (chunkSize - overlap)
- CvSearchJobTask: comment the GdprConsent = true rationale and the
BuildCvFileName sanitisation logic
- HtmlJobSearcher: comment GetLeftPart(UriPartial.Path) query-strip dedup
Closes#26
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
claude
gelu