map-wikireader / README.md
DebasishDhal99's picture
Add instructions for deployment in S3 and AWS Lambda
a724f10
metadata
title: Map-WikiReader
emoji: 🫡🥸
colorFrom: pink
colorTo: blue
sdk: docker
pinned: false

Goal

  • Combine map exploration and wiki reading.

Motivation

  • I am a map addict. While exploring Google maps, I keep a chrome tab open for checking out random places.
  • Having a handy tool to quickly look up at wiki page of random locations would be a plus.

Quick remote access

  • Use ngrok for frontend, ngrok http 3000 (Since it doesnt have the password issue)
  • Use localtunnel for backend, ngrok http 8004 (I don't want remote users to have to enter password or click on a suspicious looking link).
  • It works, I tested it on my phone, but it doesn't work on sandbox, most likely due to sandbox's requirement of Authentication from backend. Throws an Error 511.
  • Unfortunately, user still has to visit the localtunnel link to avoid Network Authentication issue. For that they would reuire the localtunnel link, and a password, which can be obtianed from https://loca.lt/mytunnelpassword, on the device that is hosting the codebase.

Resources

Deployment (Without CI/CD)

  • First of all calm down, take a backseat and relax

AWS Lambda (Backend)

  • AWS Lambda requires all the python packages and the function files, as it doesn't have the packages installed.

  • AWS Lambda architecture should have x86 architecture.

  • Make a folder lambda_build, put main.py and all the py code files keeping the structure intact (put the /backend folder i.e.)

  • Install all the files in requirements.txt (as it should be in a x86 machine).

  • To install the files, use pip install "pydantic==2.0.2" --platform manylinux2014_x86_64 --target=lambda_build --implementation cp --python-version 3.10 --only-binary=:all: --upgrade for all the libraries, handle the package name and version. (Automate it in CI/CD pipeline)

  • Once the package installation is done, time for zipping.

  • Go inside the lambda_build folder, select all files/folder, zip, upload it to Lambda function.

  • Handle the main python file. If your main file is main.py, your handler should be main.handle (if handle = Mangum(app)), these names are flexible. If main file is lamba_function.py (with lambda_handle = Mangum(app)), then handle is lambda_function.lambda_hanle.

  • Add environment files as needed on lambda. A functional url will pop up which should show {"status" : "ok"}

AWS S3 Bucket (Frontend)

  • Make a build file (npm run build)
  • S3 -> my bucket -> Properties -> Static website hosting: Enable it. Set index.html as both source and and error doc.
  • Push all the files to your S3 bucket (just the content of build, not the build folder itself)
  • Set the bucket policy as { "Version": "2012-10-17", "Statement": [{ "Sid": "PublicReadGetObject", "Effect": "Allow", "Principal": "*", "Action": "s3:GetObject", "Resource": "arn:aws:s3:::my-map-frontend/*" }] }
  • Block all public access (Turn off 'Turn off all public access`)
  • In properties, get a url to the bucket, put it inside the allowed origins of backend (Lambda)
  • That's it.