The cits3200-project from blaiddyd

Functional endpoint
Given id of a task
Return the progress of the task
- % of processed/unprocessed task
- number of animal images
- number of blank images

Tasks

Progress indicator

Description
Display the progress of an ongoing task:

%
pending count
blank count
with animal count

Definition of done

Accurately display statistics about an ongoing task

Task

Tracking
UI
Documentation
Test

File uploading form

Description
Modify the page created in #5 to include a form such that users can

Select and upload their GCP key
Select and upload multiple images

Definition of done

A form is present on the page with no errors
User can select a file on their system to upload as a key
User can only select a .json file as the key
User can select multiple images on their system to upload as images
User can only upload images that are supported by GCP Cloud Vision API

Tasks

As a user, I would like to be informed by notification when my audio transcription fails so I can fix the problem ASAP instead of waiting for the progress result.

Authentication endpoints

Description
Allow users to log in, signup and store account-specific information

Definition of done

Users can signup with
- email (unique)
- password
- name
- GCP key
User can log in using their email and password
The password isn't stored in plain text
the endpoint responds with status 200 and all of the user's information if the credentials are correct
otherwise respond with status 401

Tasks

Interface homepage

Description
A simple web page that displays the title of the module, a short description

Definition of done

Instructions on how to view the page
Able to view the page on desktops without errors
Include a sensible title and description of the module

Tasks

Create a title and a description
Implement home page
Document how to view the page

Create progress page

Description
A page that describes the status of a task

Definition of done

View progress of any task given an id
Show the % processed
Show count of animal, blank

Tasks

UI
Link to #21 endpoint
Documentation
Test

Classify cute animals

Description
Establish backend endpoint that classifies images into blank images and animals.

Definition of done

Function able to classify images given image URLs into blank or animals using the images' metadata & Vision API

Tasks

As a maintainer, I want to know how to implement a new module so I can extend the system

Write documentation on the module system
Document what each property means and how they affect the system
Documentation about the architecture of the system
Example of how one can implement a new module

As a user, I want to be able to see whether or not my video has finished processing so I know when it's finished

Front end progress page
Given a project id, show the status of the project

Animal image classification endpoint for a single image

Description
Create an HTTP endpoint that receives an image link and a GCP key (JSON object) as input and respond with whether or not that image is a photo of an animal.

Definition of done

Compiles without errors
Able to access the endpoint through localhost:{port}/{path}
Endpoint receives an image URL either through params, body or query string
Responds with status 400 if no image URL was given
Otherwise, responds with status 200 and a JSON object indicating whether the image is of an animal
Documentation on how the endpoint works
Documentation on how to use the endpoint
Accept GCP key as input

Tasks

As a user, I'd like to upload video file so I can do machine learning on them

Frontend video upload form
Similar to Eco Image Classification form
Accept video file formats
Accept JSON key

As a user, I would like to see the progress of my transcriptions so I know how I am progressing with my task.

Implement the progress function of the Audio module
Receive a project object as input
Return a ProgressReport object indicating the progress of the task
ProgressReport.done should be whether or not the task has finished (boolean)
ProgressReport.data should not be included (undefined)
When the task's resource is Pending the task is not done, when the status is Parsed it is done

Create download page

Description
A page where users can download the result of a task

Definition of done

A page that shows 2 links to download "Animal.zip" and "Blank.zip"
Clicking on either link downloads the file to their computer

Tasks

UI
Links
Documentation
Test

As a user, I want to be able to use a user friendly interface to analyse videos

Be able to access video module from the front page

As a maintainer, I want to be able to easily set up this project so I can use the system without extensive knowledge of how it works

Document the infrastructure and architecture of the project (database, object storage, etc)
Write a script that will set up the necessary infrastructure on Google Cloud Services
Document the script on what it is doing
Document the responsibility of each component of the system and how it affects everything else

As a user, I would like to download my .txt files as a .zip file so I can bulk download all my results at once.

As a user, I want to be able to transcribe an audio file that I have uploaded.

Create the process function for the audio module created in #44
Receives a project object as input
Hit the GCP Speech to Text API using the project's API key and the project's resource
Store the result in the resource's result property
The result should be a string with a newline separating every results (see doc example)

Speech to text docs

Video Object Detection

As a user, I want to identify objects in a video to quickly analyse a video

Description
Connect the form created in #7 to the endpoint created in #4 and display the result of the task to the user. Images are classified one by one and the result is immediately displayed to the user.

Definition of done

Use the image classification endpoint to process selected images
Display processed images in their respective classification
Process all images selected by the user
Process each image individually
If the endpoint fails, put the image back in the queue to try again

Tasks

Process image uploaded using the endpoint
Display processed images in different groups
Display pending images to be processed
Documentation
Test

Result download

Description
Modify the page created in #8 to allow users to download the result of the task. Users are able to download "with animal" images as a zip file and download "blank" images as another zip file

Definition of done

Users can see a download button to download the 2 groups of images
Users can download each group of images separately
Able to download each groups as a zip file

Tasks

Create Cypress Tests

Create cypress tests for end-to-end testing in order to test front-end and user experience.

Tasks

Install Cypress.
Create elementary Cypress tests.
Documentation

File downloads

Description
Implement file download endpoint and functionality in the backend. Files should be zipped for download.

Definition of done

2 separate endpoints (one for animal, one for blank)
each endpoint, given a task id, respond with a single zip file containing all images in each category

Tasks

Multiple images upload

Description
Allow multiple images to be uploaded (same logic as #4 ) and returns a response similar to #3

Definition of done

Allow multiple images to be uploaded
Responds with the classification of all images uploaded
Documentation on the process and how to use the endpoint

Tasks

Implement multiple file upload handler
documentation
test

As a user, I want to be able to get a list of objects in a video so I don't have to manually go through the file

Use Google Video Intelligence to detect objects in the videos given a project id
Use video specified in the project
Store the response of GVI back into the project on MondoDB

As a user, I want to get a list of objects as a CSV file so I can open it with other programs

Convert list of objects stored in the project record into CSV
Given a project id, return a csv file containing the objects and their timestamps
Endpoint to do the above

Bootstrap back end

Initiate backend by establishing routing and controllers.

Tasks

As a user, I want to be able to upload audio files so I can do machine learning on them.

Create a new Audio module under the audio module
make a stub for processing, progress and download (ie doesnt do anything)

Definition of done

New module appears on the front end module page
Name: Audio Transcription
Type: Audio
Clicking on the module opens the file upload form
Can upload a JSON key
Can upload 1 audio file within supportd encodings
Upload and go to progress/download page
1 type of download available called Transcription

As a user, I would like to be informed by notification when my audio task is finished so I don't have to keep checking the results page.

As a user, I want to be able to download my transcribed audio file in a text format so I can use it elsewhere.

Implement the download function of the audio module
Receive a project object
Retrieve the transcription result through the project's resource (in the result field)
Write this result to a .txt file
This function should return the absolute path to this .txt file

It is very important that the absolute path is returned and not a relative path. Have a look at other modules to see an example of path.resolve which gives you an absolute path in return.

Image uploading for image classification endpoint

Description
Instead of providing an image URL, the endpoint now accepts a single image uploaded. The endpoint should use this uploaded image to create a response instead

Definition of done

Able to use a front end web form to upload an image to the endpoint
Responds with the same status and body as #3

Tasks

Implement file upload handler
Documentation
Test

Upload Form error

Selecting files and clicking submit causes "Oh no something went wrong" error with '0/x uploaded'. Clicking 'ok' on the error message and submitting again causes the error to occur again, but with '1/x' or '2/x uploaded'. (Where x is the number of images to be processed)