User story: as a user I want to upload more files after extracting

Given: I am in upload page
When: I extract my file
Then: I will be able to upload more files

1. Create a button for uploading more files after extracting. After repeating of extracting there must be files from the previous extracting and new files too
2. Create an "extract" button, which repeats the process of extracting or showing message that nothing changed (no false positive mistakes detected and no new files to extract)

User story: as a user I want to extract all uploaded files

Create a file to extract all files from the list.

Scenario: Extract all files

Given: As a used I am in pdf page

When: I press the button “extract all”

Then: New file created, This file contains all files from the list

User story: as a user I want to write some text and see the mistakes on it

Create a field in the first page, where user can put one text. After clicking "Extract" button this text must be extracted like other pdf files

Multiple instances of the same sentence

API can send multiple instances of the same sentence
Example:

  {
    "match": "we",
    "sentence": "Second, how can we validate if we are doing a good job with online education?",
    "label": "PRONOUN4",
    "description": "Using pronoun incorrectly"
  },
  {
    "match": "we",
    "sentence": "Second, how can we validate if we are doing a good job with online education?",
    "label": "PRONOUN4",
    "description": "Using pronoun incorrectly"
  }

• the first match corresponds to first 'we' in the sentence that starts at character 17
• the second match corresponds to second 'we' in the sentence that starts at character 32

Our application responds to this type of sentences

Our app just search for match (we) in a sentence and mark all of them as mistake phrases. The API sends this as two separate sentences that why our application shows it twice.

Possible solutions:

1. Create a map Map<Object, bool> to store object with match, sentence, label, description fields, if we receive some sentence from API save it(make it true) and if we'll receive the same(if it already true) object ignore it.
   • It will solve this bug
   • We'll consume extra memory to do that
2. If we knew the position(index) of incoming match this bug would be solved easily

Other Bugs

Will be discussed two types of sentence:

1. match = 'n't'
2. match = 'we'

API also can send this kind of sentence:

  {
    "match": "n't",
    "sentence": "Lastly, students like to ask questions right away, and they don't mind if their colleagues answered the questions instead of the teacher.",
    "label": "SPOKN1",
    "description": "Using contractions"
  }

Our application just searching for match (n't) that we received from API
The search process for our program:

  while (true) {
    var index = text.indexOf(match, start);
    ...
    // using index to get match in sentence
   ...
   start = index + (match.length as int);
  }

Our application responds to this type of sentences

It looks like our application is working correctly, but the search approach can cause the following problem:

API also can send this kind of sentence:

  {
    "match": "we",
    "sentence": "To answer our research questions first we decided to collect data through a survey form.",
    "label": "PRONOUN4",
    "description": "Using pronoun incorrectly"
  }

Our application responds to this type of sentences

After all, it looks like there are interesting bugs in our application:grin:

so the search approach described above is not perfect

Possible solution for second type:

1. mark as mistake only phrases that has space from left and right (if match='we' then in answer phrase 'we' won't be considered as mistake)

But this approach can cause bugs in sentences of the first type(where match='n't')

Possible solution for both types:

1. If we knew the position(index) of incoming match this bug would be solved easily
2. in progress:sweat_smile:

Implement the solution that was described for specific bugs.

User story 1: as a user I want to delete some files after uploading

Given: I am in upload file page
When: I uploaded a couple of file
Then: If I don't need any of files I could delete them by clicking some button

Create a sign near all items in the list of extracted files. After clicking this sign file must be removed from the list

User story: as a user I want to look at user-friendly interface

Create a design for the page with extracted files:

• The page must be more beautiful
• The page must have user friendly functionality
• The tutle must be a link to the main page

User story: as a user I want to see the list of mistakes for all extracted files

Create a function to parse files for using the material from them to API
The function must separate each file and get as a result csv files with mistakes for the one file only.
Also the function must create csv file with mistakes from all files together

User story: as a user I want to detect false positive mistakes

Given: I uploaded a file
When: I am looking at mistakes that API gave to me
Then: If provided mistake sentence is actually wrong, then i could send feedback to API about that

Create a sign near each mistake. After clicking this sign the mistake must be removed from the list and marked as false positive mistake (for sending info about that to API creator if necessary)

Create page, where there is a button to upload pdf files. Uploaded files must be visible in the page

As a user, I want to upload some files to extract it

Create page, where there is button "Upload". By clicking this button user can upload some pdf files. There is also a list of uploaded files in the page.

User story: as a user I want to fix one file to see the mistakes

• Create a button to extract only one file from the list
• Create a button to extract the text from the text field
  Scenario: Extract one file

Given: As a user, I want to be able to fix one file, and the see mistakes
When: I press the fix button,
Then: currently selected file extracted from the list, button which allows me to extract the text from the file appears, file removed from the list.

User story: as a user I want to see the list of mistakes for all extracted files

Create a field for one mistake. Each mistake should contain:

• the mistake itself
• A sentence with highlighted place, where the mistake is
• A description of the mistake (how to fix or what the problem is)

Scenario: create one mistake

Given: As a user I want to see the list of mistakes for all extracted files

When: Files are extracted
Then: I see the list of mistakes, where each mistake contain: the mistake itself, a sentense with highlighted place, a description of the mistake

User story: as a user I want to look at user-friendly interface

Improve the design of the page, where there is text field

• The design must be more high colorful
• The text field must be smaller
• The buttons must be replaced