gg_fnc_assignment

Features:

• When an image is uploaded to the firebase storage. the function processImage will be trigger and analyse the content inside the image. It will print out the label of that image. Example:

 [ { locations: [],
     properties: [],
     mid: '/m/07s6nbt',
     locale: '',
     description: 'Text',
     score: 0.9882610440254211,
     confidence: 0,
     topicality: 0.9882610440254211,
     boundingPoly: null },
   { locations: [],
     properties: [],
     mid: '/m/03gq5hm',
     locale: '',
     description: 'Font',
     score: 0.9290081262588501,
     confidence: 0,
     topicality: 0.9290081262588501,
     boundingPoly: null },
   { locations: [],
     properties: [],
     mid: '/m/03scnj',
     locale: '',
     description: 'Line',
     score: 0.8226431012153625,
     confidence: 0,
     topicality: 0.8226431012153625,
     boundingPoly: null } ] 

What problems did I face when building this function?

• At the time I received this assignment, I have no experience about Google cloud function and firebase storage.
• I have started to search in the internet what are Cloud functions and Firebase Storage
• When I started to implement it. I faced with different versions of firebase sdk, leads to different method and function name.
• I have spend about 3 hours for the basic function renamed image (when someone upload an image, this function will trigger just for rename (just for studying purpose)). Of course, It did not take too much time like that. I had a problem with the renamed image, which can not be public automatically at the beginning. So that, I have to completely find out the approach for generate access token for the renamed image.
• After finish my own small challenge. I went back to the task.
• I found a service in Google Cloud, named Cloud Vision API that allows developers can execute some image processing and other AI & ML stuffs (this is my first task relate to Machine Learning)
• I have recognized that my current node version in this project is 8. So that I have to upgrade it to 10 in order to use Cloud Vision API
• I have finished this task in the morning of Oct 7th 2020. However, I still want to find out the solution for reuse it later.

Improvement?

• Well, I think for now, this function is not so helpful (we can only preview it in log tab). It should store the result data in database for using later.
• I think we could use http cloud function for upload image, then also execute image processing, and return the json result