mosquito-alert / mosquito_alert Goto Github PK

This has several dependencies to be aware of: map, map publication and entolab

is not flagged

in case of report: + is_public

Fix this:

https://github.com/MoveLab/tigatrapp-server/blob/master/tigaserver_app/models.py#L549

It's already corrected in the repository/master, but latest version is not yet in production

Scenario: we want to get all boundaries that contain a point.

Current strategy:
Using a queryset that queries all boundaries in DB and get those that contains the desired point.

Proposed strategy:
Query first only boundaries from the highest boundary_level (aka: countries), which contains less objects, and get the/those boundaries that contain the point. Then, query the boundaries from the next boundary_level and filter only that are children from the previous result... And so on until reaching the last boundary_level

Candidates:
https://github.com/abhiabhi94/django-flag-app
https://github.com/idlesign/django-siteflags
https://github.com/caffeinehit/django-flaggit
https://github.com/pinax/django-flag
https://github.com/dominno/django-moderation/

https://django-contrib-comments.readthedocs.io/en/latest/moderation.html

Example:
Rank Specie:
OK -> Aedes albopictus
KO -> Aedes Albopictus

Rank other than species: allow anything

We are trying to consolidate all of the Mosquito Alert repositories in the Mosquito-Alert organization. This is low-priority, but at some point we should transfer this one after double-checking that doing so will not break anything. My experience so far is that this is very easy: Just transferring ownership under Settings will move the repository and also create redirects such that all of the existing URLs continue to work (both the GithHub URLs for anyone clicking an old link to the MoveLab/tigaserver repository, and the git remotes on our local machines and servers for pulling and pushing code).

Individual is in a photo inside a boundig box
ComputerVision return classes in a bounding box

https://docs.aws.amazon.com/rekognition/latest/dg/images-displaying-bounding-boxes.html

See: https://waf.cs.illinois.edu/visualizations/Perception-of-Probability-Words/

Thanks @UkoZJ

Example:

polygon = Polygon.from_bbox(bbox=(0,0,10,10))
polygon_children = Polygon.from_bbox(bbox=(1,1,5,5))

point_in_children = Point(2,2)

boundary_a = Boundary(geometry=polygon)
boundary_a_child = Boundary(parent=boundary_a, geometry=polygon_children)

loc = Location(point=point_in_children)

assert frozenset(loc.boundaries.all()) == frozenset([boundary_a, boundary_a_child])

# Unlink parent boundary
boundary_a.locations.remove(loc)

# HERE: child should have been removed to

assert loc.boundaries.all() == []

LINKING PARENT -> SHOULD TRY TO LINK ALL THE DESCENDANTS
UNLINKING PARENT -> UNLINK ALL THE DESCENDANTS

LINKING CHILDREN -> MUST LINK ALL PARENTS
UNLINKING CHILDREN -> Do nothing

Shareable means:
- is_public
- license allows being shared

to:

• photos
• reports

add attributes

add is_transition in monthly distribution?

Apart from import_export, take https://github.com/tolomea/django-data-browser into account

Decide how to trigger that bool

We don't need to obtain precision on distance computation, only the geolocation of points. geometry will be faster. srid=4326

https://arxiv.org/pdf/1903.02050.pdf
https://github.com/yding5/AdaptiveBinning

Thanks to @UkoZJ

In order to exchange occurrence data, Darwin Core offers an XML interchange format.

https://dwc.tdwg.org/simple/

We should create an API that enables other systems to harvest records through Darwin Core (with the proper authentication method).

See also https://www.gbif.org/infrastructure/registry since everything we create should be GBIF-harvestable as well if necessary.

Show URL in admin

Decide if report -> individual is:

• 1:1
• 1:N
• N:1

A user can have a role in the organization.
Users with certain roles can invite others
Output (products) are linked to the organization.
Users can have a role in a team/organization.

Different types:

• Town halls
• Public health agencies
• etc…

See: https://django-organizations.readthedocs.io/en/latest/

Photo is rotated, but EXIF still have the Orientation tag wrong.

• renovate (#96)
• ci: pre-commit.ci + pytest
• build docker + push in ghcr (https://docs.github.com/en/packages/managing-github-packages-using-github-actions-workflows/publishing-and-installing-a-package-with-github-actions)
• coverage

Candidates:
https://pypi.org/project/django-postman/
https://github.com/abdurraufraihan/django-chat/tree/main/server
https://github.com/Madhur215/Django-ChatApp/tree/master

Check: https://github.com/epicserve/django-base-site

Determine a threshold of "good quality"

Data source candidates:

• https://pypi.org/project/suntime/
• https://pypi.org/project/suntimes/
• https://pypi.org/project/sunriset/
• https://pypi.org/project/astral/

Benchmarks:

• https://stackoverflow.com/questions/38986527/sunrise-and-sunset-time-in-python
• https://stackoverflow.com/a/70186517/11671779

Conclusion: use https://rhodesmill.org/skyfield/

Take into account places where the sun never reaches certain degrees below the horizon (e.g: northern Sweden)

Report type: IndividualReport & BiteReport

ACHTUNG: Use local datetime (TZ) as reference

Interesting info:

• Daylight duration
• Sun elevation
• Category
• Sunset.datetime
• Diff between created time and sunset event
  • report.datetime - sunset.datetime (or NULL, if the sun is never below 6deg that day, e.g: Northern Norway)
  • after 00h00 difference with tomorrow.sunset.datetime
• moon phase

To be read: https://besjournals.onlinelibrary.wiley.com/doi/10.1111/2041-210X.13290

Adapt the current superexpert revision capability to allow national supervisors to revise the reports validated by IA. The national supervisor executively validates the report and removes the IA validation.

Start transaction.atomic when objects are saved in db.
Check if transaction.atomic is needed... Not sure if import_export is already guaranteeing it

Ensure transaction lock is as short as possible

If a user creates and identificationsuggestion:

• Create identification suggestion
• Mark individual as reviewed by that user

If a user doesn't how to identify the individual and want to forget about it:

• Mark it as reviewed + do not show

Users have to be able to annotate certain characteristics from individuals such as:

• stage of life
• sex
• does it has blood?
  etc

Hard voting system

Should be implemented with push