Spaces:
Runtime error
Runtime error
Aryan Mishra commited on
Commit Β·
4347000
1
Parent(s): 6bf2e25
Remove CI/CD references from docs
Browse filesDeleted the GitHub Actions workflow files and updated the README and project docs to remove outdated CI/CD mentions, keeping the repository documentation aligned with the current deployment and monitoring setup.
- .github/workflows/ci.yml +0 -38
- .github/workflows/deploy.yml +0 -29
- .github/workflows/drift_check.yml +0 -20
- README.md +0 -1
- docs/architecture.md +0 -12
- docs/demo_script.md +2 -5
- docs/resume_bullets.md +3 -3
.github/workflows/ci.yml
DELETED
|
@@ -1,38 +0,0 @@
|
|
| 1 |
-
name: CI
|
| 2 |
-
on: [push, pull_request]
|
| 3 |
-
jobs:
|
| 4 |
-
test:
|
| 5 |
-
runs-on: ubuntu-latest
|
| 6 |
-
steps:
|
| 7 |
-
- uses: actions/checkout@v4
|
| 8 |
-
- uses: actions/setup-python@v5
|
| 9 |
-
with: {python-version: "3.11"}
|
| 10 |
-
- run: pip install -r requirements.txt mypy
|
| 11 |
-
- run: PYTHONPATH=. pytest tests/ -v --tb=short
|
| 12 |
-
- run: PYTHONPATH=. python -m mypy src/ --ignore-missing-imports
|
| 13 |
-
|
| 14 |
-
lint:
|
| 15 |
-
runs-on: ubuntu-latest
|
| 16 |
-
steps:
|
| 17 |
-
- uses: actions/checkout@v4
|
| 18 |
-
- run: pip install ruff black
|
| 19 |
-
- run: ruff check src/ api/
|
| 20 |
-
- run: black --check src/ api/
|
| 21 |
-
|
| 22 |
-
build-docker:
|
| 23 |
-
needs: [test, lint]
|
| 24 |
-
runs-on: ubuntu-latest
|
| 25 |
-
steps:
|
| 26 |
-
- uses: actions/checkout@v4
|
| 27 |
-
- uses: docker/setup-buildx-action@v3
|
| 28 |
-
- uses: docker/login-action@v3
|
| 29 |
-
with:
|
| 30 |
-
registry: ghcr.io
|
| 31 |
-
username: ${{github.actor}}
|
| 32 |
-
password: ${{secrets.GITHUB_TOKEN}}
|
| 33 |
-
- uses: docker/build-push-action@v5
|
| 34 |
-
with:
|
| 35 |
-
context: .
|
| 36 |
-
file: config/docker/Dockerfile.api.prod
|
| 37 |
-
push: ${{github.ref == 'refs/heads/main'}}
|
| 38 |
-
tags: ghcr.io/${{github.repository}}/api:latest
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
.github/workflows/deploy.yml
DELETED
|
@@ -1,29 +0,0 @@
|
|
| 1 |
-
name: Deploy
|
| 2 |
-
on:
|
| 3 |
-
push:
|
| 4 |
-
branches: [main]
|
| 5 |
-
jobs:
|
| 6 |
-
deploy-api:
|
| 7 |
-
runs-on: ubuntu-latest
|
| 8 |
-
steps:
|
| 9 |
-
- uses: actions/checkout@v4
|
| 10 |
-
- name: Deploy to Railway
|
| 11 |
-
run: |
|
| 12 |
-
npm install -g @railway/cli
|
| 13 |
-
railway up --service api
|
| 14 |
-
env:
|
| 15 |
-
RAILWAY_TOKEN: ${{secrets.RAILWAY_TOKEN}}
|
| 16 |
-
|
| 17 |
-
deploy-dashboard:
|
| 18 |
-
runs-on: ubuntu-latest
|
| 19 |
-
steps:
|
| 20 |
-
- uses: actions/checkout@v4
|
| 21 |
-
- uses: actions/setup-node@v4
|
| 22 |
-
with: {node-version: "20"}
|
| 23 |
-
- run: cd dashboard && npm install && npm run build
|
| 24 |
-
- uses: amondnet/vercel-action@v25
|
| 25 |
-
with:
|
| 26 |
-
vercel-token: ${{secrets.VERCEL_TOKEN}}
|
| 27 |
-
vercel-org-id: ${{secrets.VERCEL_ORG_ID}}
|
| 28 |
-
vercel-project-id: ${{secrets.VERCEL_PROJECT_ID}}
|
| 29 |
-
working-directory: dashboard
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
.github/workflows/drift_check.yml
DELETED
|
@@ -1,20 +0,0 @@
|
|
| 1 |
-
name: Weekly Drift Check
|
| 2 |
-
on:
|
| 3 |
-
schedule:
|
| 4 |
-
- cron: "0 9 * * 1" # Every Monday 9am
|
| 5 |
-
jobs:
|
| 6 |
-
drift-check:
|
| 7 |
-
runs-on: ubuntu-latest
|
| 8 |
-
steps:
|
| 9 |
-
- uses: actions/checkout@v4
|
| 10 |
-
- uses: actions/setup-python@v5
|
| 11 |
-
with: {python-version: "3.11"}
|
| 12 |
-
- run: pip install -r requirements.txt
|
| 13 |
-
- run: python scripts/drift_monitor.py
|
| 14 |
-
env:
|
| 15 |
-
DATABASE_URL: ${{secrets.PROD_DATABASE_URL}}
|
| 16 |
-
MLFLOW_TRACKING_URI: ${{secrets.MLFLOW_TRACKING_URI}}
|
| 17 |
-
- uses: actions/upload-artifact@v4
|
| 18 |
-
with:
|
| 19 |
-
name: drift-report
|
| 20 |
-
path: monitoring/reports/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
README.md
CHANGED
|
@@ -33,7 +33,6 @@ This repository is designed following a **Python-first paradigm**:
|
|
| 33 |
|
| 34 |
```text
|
| 35 |
Multilingual-Absa/
|
| 36 |
-
βββ .github/workflows/ # CI/CD pipelines
|
| 37 |
βββ api/ # FastAPI backend and inference services
|
| 38 |
βββ config/ # DVC and Docker configuration files
|
| 39 |
βββ dashboard/ # React frontend for inference & monitoring
|
|
|
|
| 33 |
|
| 34 |
```text
|
| 35 |
Multilingual-Absa/
|
|
|
|
| 36 |
βββ api/ # FastAPI backend and inference services
|
| 37 |
βββ config/ # DVC and Docker configuration files
|
| 38 |
βββ dashboard/ # React frontend for inference & monitoring
|
docs/architecture.md
CHANGED
|
@@ -28,15 +28,3 @@ graph LR
|
|
| 28 |
F --> G[Stage 2: Sentiment Classifier]
|
| 29 |
G --> H[aspect, sentiment, confidence]
|
| 30 |
```
|
| 31 |
-
|
| 32 |
-
## 3. CI/CD Pipeline
|
| 33 |
-
```mermaid
|
| 34 |
-
graph LR
|
| 35 |
-
A[git push] --> B[GitHub Actions]
|
| 36 |
-
B --> C[pytest + ruff]
|
| 37 |
-
C --> D[Docker build]
|
| 38 |
-
D --> E[Push to GHCR]
|
| 39 |
-
E --> F[Deploy to Railway]
|
| 40 |
-
B --> G[npm build]
|
| 41 |
-
G --> H[Deploy to Vercel]
|
| 42 |
-
```
|
|
|
|
| 28 |
F --> G[Stage 2: Sentiment Classifier]
|
| 29 |
G --> H[aspect, sentiment, confidence]
|
| 30 |
```
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
docs/demo_script.md
CHANGED
|
@@ -23,10 +23,7 @@
|
|
| 23 |
*Action*: Switch tabs to the Grafana dashboard (`localhost:3001` or deployed URL).
|
| 24 |
*Narration*: "Production reliability is crucial. Here we see our Prometheus metrics scraped from FastAPI: request rates, P95 latency consistently under 200ms, and Celery worker health."
|
| 25 |
|
| 26 |
-
**02:45 β Show
|
| 27 |
-
*Action*: Switch tabs to the GitHub Actions page.
|
| 28 |
-
*Narration*: "Every push goes through a rigorous CI/CD pipelineβtesting, linting, Docker builds, and automated deployments to Railway and Vercel."
|
| 29 |
-
|
| 30 |
-
**03:00 β Show HuggingFace Hub model page**
|
| 31 |
*Action*: Switch to the HuggingFace Hub repository.
|
| 32 |
*Narration*: "Finally, our optimized INT8 ONNX models are hosted publicly on HuggingFace Hub. Thank you for watching!"
|
|
|
|
|
|
| 23 |
*Action*: Switch tabs to the Grafana dashboard (`localhost:3001` or deployed URL).
|
| 24 |
*Narration*: "Production reliability is crucial. Here we see our Prometheus metrics scraped from FastAPI: request rates, P95 latency consistently under 200ms, and Celery worker health."
|
| 25 |
|
| 26 |
+
**02:45 β Show HuggingFace Hub model page**
|
|
|
|
|
|
|
|
|
|
|
|
|
| 27 |
*Action*: Switch to the HuggingFace Hub repository.
|
| 28 |
*Narration*: "Finally, our optimized INT8 ONNX models are hosted publicly on HuggingFace Hub. Thank you for watching!"
|
| 29 |
+
|
docs/resume_bullets.md
CHANGED
|
@@ -4,11 +4,11 @@
|
|
| 4 |
> Multilingual ABSA system: XLM-RoBERTa fine-tuned for aspect-level sentiment in English/Hindi, deployed with <300ms latency.
|
| 5 |
|
| 6 |
**Project bullet (for projects section):**
|
| 7 |
-
> Built end-to-end Multilingual ABSA platform supporting English and Hindi: fine-tuned XLM-RoBERTa achieving 78.1% EN / 67.8% HI macro-F1; exported to ONNX INT8 for <300ms CPU inference; deployed FastAPI backend on Railway + React dashboard on Vercel with
|
| 8 |
|
| 9 |
**Skills demonstrated (for recruiter talking points):**
|
| 10 |
- NLP: ABSA, BIO tagging, multilingual transfer learning
|
| 11 |
- MLOps: MLflow, DVC, Evidently AI, drift monitoring
|
| 12 |
- Engineering: FastAPI, Celery, Docker, PostgreSQL, Redis
|
| 13 |
-
- Deployment: Railway, Vercel, HuggingFace Hub
|
| 14 |
-
- Monitoring: Prometheus, Grafana
|
|
|
|
| 4 |
> Multilingual ABSA system: XLM-RoBERTa fine-tuned for aspect-level sentiment in English/Hindi, deployed with <300ms latency.
|
| 5 |
|
| 6 |
**Project bullet (for projects section):**
|
| 7 |
+
> Built end-to-end Multilingual ABSA platform supporting English and Hindi: fine-tuned XLM-RoBERTa achieving 78.1% EN / 67.8% HI macro-F1; exported to ONNX INT8 for <300ms CPU inference; deployed FastAPI backend on Railway + React dashboard on Vercel with Prometheus monitoring, and weekly drift detection.
|
| 8 |
|
| 9 |
**Skills demonstrated (for recruiter talking points):**
|
| 10 |
- NLP: ABSA, BIO tagging, multilingual transfer learning
|
| 11 |
- MLOps: MLflow, DVC, Evidently AI, drift monitoring
|
| 12 |
- Engineering: FastAPI, Celery, Docker, PostgreSQL, Redis
|
| 13 |
+
- Deployment: Railway, Vercel, HuggingFace Hub
|
| 14 |
+
- Monitoring: Prometheus, Grafana
|