geotiff / geoblaze Goto Github PK

http://ec.europa.eu/eurostat/web/gisco/geodata/reference-data/population-distribution-demography/geostat

already did sum, have to do others, too

ift.tt/2gKrEzF

Params:

Clip By This Raster:

• raster loaded
• values (exact or range)

Raster To Clip

(No Data Value? or take from GeoTIFF metadata)

I think it'll be faster if we have an if statement or ternary expression in the reduce function rather than having filter then reduce. That's because we're currently iterating over all the values twice. Not sure though. We can write performance tests too to find out!

@DanielJDufour DanielJDufour Merge pull request #24 from sdpeyton/temp …

Should probably rename to something a little more intuitive

With features getting more complex, testing will need to be re-architected to be more efficient at getting high coverage. Specifically, we need:

• a common access point for geometries
• coverage of bbox, single and multi-part geometries, feature collections, union geometries
• coverage for all possible geometry input formats
• a standard for how these tests are asserted

Currently, tools check to see if there is only one band and returns just those results if that is the case. This will be unsustainable in the future; it is better to have the user handle it on the front end

Need to write tests (probably in a separate repository) that tests and compares GDAL, GeoDjango, rasterio, and geoblaze.

We want to run automated tests after each PR created.

You want to place a .travis.yml file in the root directory and configure it to run npm test. Here's an example for a Python package.

https://github.com/DanielJDufour/date-extractor/blob/master/.travis.yml

https://travis-ci.org/

Identify runs slow on mobile devices probably because readRasters takes a while to run (not 100% sure on this though).

We could speed this up by doing the following:

• converting the values stored in the geotiff.js object to a regular array when the size of the object isn't too big. This would mean that you load the values into JavaScript and when you run an identify, you directly grab the element in the array that corresponds with the user-designated cell.

We can use the Browser's performance API, to track how long operations take.

which are what the Load File option uses

Here's an example of a raster that uses a float for their GDAL_NO_DATA value

It's a Rwanda raster for world pop.

https://github.com/GeoTIFF/geotiff.io/blob/master/data/RWA_ppp_v2b_2020.tif

using documentation.js that MapBox GL JS and Turf use

Here's example of geojson:
http://bl.ocks.org/d/17662f9239fd6e4c38008c48125af756

{
"type": "FeatureCollection",
"features": [
{
"type": "Feature",
"properties": {},
"geometry": {
"type": "Polygon",
"coordinates": [
[
[
-112.1484375,
-55.37911044801048
],
[
101.6015625,
-55.37911044801048
],
[
101.6015625,
68.9110048456202
],
[
-112.1484375,
68.9110048456202
],
[
-112.1484375,
-55.37911044801048
]
]
]
}
}
]
}

need to return row index and column index along with value and band_index.

It'd be nice of the Readme had the following categories, if not more:

• Basic Description
• Installation npm install geotiff-io
• Usage Examples
• Table of high-level features (e.g., identify, viewshed, etc.)
• How To run the test
• Also link to documentation

Here's an example:
https://github.com/DanielJDufour/date-extractor

Is your feature request related to a problem? Please describe.

We need to set up automated testing, so we can automatically test pull requests

Describe the solution you'd like

We probably should create a .circleci folder and create a config.yml inside of this folder. This config.yml configures the Circle CI tests. Here's an example of a config.yml file: https://github.com/FirstDraftGIS/cooccurrences/blob/master/.circleci/config.yml

Describe alternatives you've considered

TravisCI also provides from testing for Open Source projects. I'd be happy with Travis CI, too! :-) Here's an example of a .travis.yml file: https://github.com/DanielJDufour/date-extractor/blob/master/.travis.yml

Additional context

N/A

Here's the link to the file:

ftp://ftp.glcf.umd.edu/glcf/SRTM/Degree_Tiles/n038/SRTM_f03_n038e077/SRTM_f03_n038e077.tif.gz

check out UTIF
https://github.com/photopea/UTIF.js?utm_source=tuicool&utm_medium=referral

• Should we add code editor to docs, so people can actually edit example code on the fly???

• We probably should only have a few rasters loaded for all the samples though...

https://github.com/ajaxorg/ace

Is your feature request related to a problem? Please describe.
A user should be able to clip a raster by a vector. In other words, the clip function should clip the raster by the geometry's bounding box and set all the remaining values outside of the geometry to No Data. It should also update the metadata for the raster.

Describe the solution you'd like
A clear and concise description of what you want to happen.
Create a clip module that takes in a georaster and geovector and returns the result as a georaster. The approach could include:

1. Calculating bbox of geometry
2. Getting insides of the geometry
3. Iterating through rows of pixels, creating a new row starting at xmin of geometry, and adding NoData where outside geometry and current pixel value where inside the geometry

Describe alternatives you've considered
None at the moment

Additional context
You'll need to borrow some logic from https://github.com/GeoTIFF/geoblaze/blob/master/src/intersect-polygon/intersect-polygon.module.js and potentially split out some of the code, creating a separate getInsides module

gio-convert-geometry should handle points when passed in as object with x and y as properties

Since it is the standard convention in javascript, we should more to camel case.

I think it would be awesome if we could create ObservableHQ Notebook examples of raster band math. It'd also help users understand how to use it. You could also push it to app.geotiff.io like I do in https://beta.observablehq.com/@danieljdufour/create-an-rgb-geotiff if the GeoTIFF file isn't too big to embed in the url!

Documentation https://docs.geoblaze.io/ looks ugly (I did it)

We need to make it look better, something like https://turfjs.org

Perhaps we can borrow orange color from http://geoblaze.io/

Currently, documentation doesn't describe that third options parameter for histogram. A user spent a long time and had to go through our source code to figure out what params to pass in with options. We need to update documentation.

use namespaced

Throw errors in gio.js. Could you throw an errors in addition to logging to console with console.error? We need this in order to capture errors and alert the users.

We can then catch errors in promises with .catch (the opposite of .then).

https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Promise/catch

We should add a debug mode in the settings, also as a url param, that will add a lot of alerts to the page with relevant info. For example, after we run an identify, the alert should say how much each part of that process took.

We absolutely don't want to load all of the raster pixel values into JavaScript memory for a simple identify function. We should use a rasterWindow, which you can be seen used here: https://www.npmjs.com/package/geotiff

Using RasterWindow will also decrease the odds of maxing out the system memory and crashing the browser.

Not necessarily gonna rearchitect

Code complexity will become an issue as we increase the number of features we support. Converting to classes may improve readability

This is a research task - if this refactor would significantly hurt performance, we should abort