kobotoolbox / formpack Goto Github PK

For each attachment (video, image, recording, audit.csv, etc.) that was sent with a submission, include the URL to the file in an additional column named [original_question_name]_URL.

For example

Related to #43 and #46

At a minimum, allow an optional 2nd header row of an export to contain #hxl tagged columns

Currently all number responses in the Excel export are saved as text. Instead they should be saved as numbers (to allow the user to run calculations without having to manually change the data type each time).

Needs a fixture + test

Currently all files are just named by the form ID, which is not helpful in organizing many exported files.

[form title] - [type] - YYYY-MM-DD - HH-MM-SS.[extension]

Whereas [type] can be:

• values
• labels
• language1
• language2

• select_multiple is not split into different columns
• responses are not displayed using language labels
• when using 'label' and 'label::language' only 'label::language' is available in ktable
• uuid, submision time not included in exports
• first select_one question missing from XLS export. See
  screenshot
• repeat group sheet in XLS doesn't include responses for one question (and the question is shown in a weird place) See screenshot
• XLS exports can't open in Excel
• gps is not split into multiple columns

(otherwise leading zeros are stripped from e.g. phone numbers, ID numbers)

Select specific response, e.g. 'gender = female'

Additional filters will be done in a future story.

• for blank responses in select_multiple questions, don't write 0 and 1 in the split columns (it's a null response and these need to stay empty as well)
• text and barcode response columns should be formatted as text in Excel (otherwise leading zeros are stripped from e.g. phone numbers, ID numbers)
• integers are stored as text in Excel, they should be stored as General or Numbers
• Dates are not recognized in Excel (en_US)
• the order of columns for split select_multiple questions is not correct: it should be the same order as in the form (right now seems to be random)
• parent groups are duplicated in split GPS columns

original question: d/G_A/group_qb6pp39/GPS_coordinates
Expected split question: d/G_A/group_qb6pp39/GPS_coordinates_latitude
Current: d/G_A/group_qb6pp39/_d/G_A/group_qb6pp39/GPS_coordinates_latitude

• leading and trailing spaces from Text responses are not included in cells but should be (like in CSV)
• _id column is missing from export (kc's pk), this may be important to some power users

Fixtures of forms and data featuring:

• Skip logic
• Groups
• Translations
• Loops
• Versioned forms

For each media attachment (video, image, recording) that was sent with a submission, include a direct link to the file in an additional column named [original_question_name] - URL.

This URL should be a clickable link, which can be achieved with the 'hyperlink' function in Excel. (But not necessary as long as URL is interpreted as a link by Excel).

=hyperlink("https://kc.kobotoolbox.org/attachment/original?media_file=[username]/attachments/[filename]")

Related to #43.

We need the ability to import from and export to a number of different formats.

• import ss[json]
• import xls
• import instance xml
• import instance json
• import xml[zip]
• import from formpack[json]
• import from django queryset
• export to formpack[json]
• export to xml[zip]
• export to photos[zip]

implementation detail

Dates are not recognized in Excel

yyyy-mm-dd isn't recognized by Excel (see current KC date exports which works in all locales)

For each media attachment (video, image, recording) that was sent with a submission, include a direct link to the file:

<a href="https://kc.kobotoolbox.org/attachment/original?media_file=[username]/attachments/[filename]">[filename]</a>

Details of how the columns and rows should be organized are yet to be specified

It's quite difficult to debug problems of malformed input. Fixing this will save us time in the long run.

include a way to export a properly formatted OrderedDict to an XLS file with ordered named sheets.

Also, consider that the language of the table must be configurable

TODO: Create an example of the data exports for each data type

the order of columns for split select_multiple questions is not correct: it should be the same order as in the form (right now seems to be random). When a select_multiple response wasn't ticked in any submission it should still be included as a blank column in the dataset. Formpack currently excludes it.

See screenshot:

Expected output would be this (column b follows column a, as in the form design):

Important for pagination.

Display submissions 1-10, 11-20, etc.

Example of a "region" disaggregated export to be pasted below