-
geeView UI Layout Enhancements - UI has had minor enhancements to improve use of space and alignment.
-
upload_to_gcs overwrite setting -
assetManagerLib.upload_to_gcsoverwiteparameter. By default, it is now set toFalse. This can be a breaking change since before it was unset, leaving it to essentially beoverwrite = True. -
cloudStorageManager new functions -
cloudStorageManager.list_files,cloudStorageManager.bucket_exists, andcloudStorageManager.create_bucketfunctions to facilitate various cloud storage tasks. -
geeView query yLabel setting - Can now set the
yLabel(vizParams["queryParams"]["yLabel"]) to label the y axis of a query chart. See pydoc forMap.addLayerandMap.addTimeLapsefor details. -
geeView minZoomSpecifiedScale - Can now set min zoom level to start changing zonal stats reduction scale changes (
vizParams["areaChartParams"]["minZoomSpecifiedScale"]). This is useful to avoid memory errors, but ideally, can be set to a lower zoom level if possible. See pydoc forMap.addLayerandMap.addTimeLapsefor details.
-
Doc updates - Continued work improving Documentation website. Numerous function-wise examples added to the code.
-
GEE Palettes Integration - Converted Gena's EE Palettes json to a Python script/module. Use the module as
import geeViz.geePalettes as palettes
Map.addLayerandMap.addTimeLapsefunctions now create a copy of thevizparam dictionary. If you passed a dictionary, and then reused that in another layer call, thelayerTypekey was then populated and wouldn't be updated, this resulting in an object type error. By creating a copy, the dictionary isn't changed outside the function.
-
Doc updates - Continued work improving Documentation website.
-
Simple Sentinel-2 -
getImagesLib.superSimpleGetS2is a super lightweight function to get analysis-ready Sentinel-2 data integrating the cloudScore+ algorithm. This may be a better option than existing functions that support out-dated legacy methods that are no longer needed since cloudScore+ works so well.
-
Doc updates - Continued work improving Documentation website.
-
MapBiomas Example Rework - The MapBiomas example script and notebook now integrate a full mosaic of all study areas and the ability to collapse to lower levels of the classification hiearchy.
- Doc updates - Continued work improving Documentation website - logo bug fix..
-
Doc updates - Continued work improving Documentation website.
-
New and Updated Example Notebooks - Global Landcover notebook added and updates to MapBiomas example notebook and script.
- Doc updates - Continued work improving Documentation website. Notably integrating notebook examples into doc website.
- Function parameter reordering bug fix for
getImagesLib.getS2
-
CloudScore+ cs_cdf band support - You can now utilize the
"cs_cdf"band from the cloudScore+ algorithm in any of the Sentinel-2 image prep functions in thegetImagesLib(getS2,getProcessedSentinel2Scenes,getProcessedLandsatAndSentinel2Scenes,getSentinel2Wrapper,getLandsatAndSentinel2HybridWrapper) by specifyingcloudScorePlusScore = "cs_cdf". -
LayerType optimizing - The specification of the
layerTypeinsize the visualization pareters in aMap.addLayerfunction is now largely unnecessary since an efficient object type method has been added.
- Area charting add layer bug fix for adding area chart layers without delcared object type.
-
Larger Area Charts for Downloading - Area charts no automatically scale by 2x for downloading
-
Sankey Charts Filtering - Any class less than a specified percentage (
sankeyMinPercentage- default :0.5%) won't be shown in the sankey chart. This helps clean up largely meaningless transitions from being shown. -
Composite uploading support -
assetManagerLib.uploadToGEEImageAssetnow supports composite uploading using the"GSUtil:parallel_composite_upload_thresholdsetting.
- Sentinel-2 QA bands being set to null instead of 0 since ~Feb 2024 bug fixed. QA bands are no longer required to have values for a pixel to be valid.
- Google Colab bug fix - earthengine-api changed
ee.oauth._in_colab_shell()toee.oauth.in_colab_shell(). Backward compatability supported for older function name.
- Area charting bug fixes and time lapse uneven dates slider bug fix.
- Area Charting Example Notebook - There is now an area chart example notebook that walks through many examples of the available methods for charting zonal stats of ee images and imageCollections.
- Area charting bug fixes for various data types including thematic images.
-
You can now use
Map.addLayerorMap.addTimeLapsewith an image collection where"canAreaChart" : Trueand"reducer : ee.Reducer.frequencyHistogram()but nobandName_class_values,bandName_class_names,bandName_class_paletteare provided. Previously, this would not work. -
Fixed bug where
areaChartParamsreducerwas not being handled properly forMap.addTimeLapsecalls.
-
Documentation for geeViz - We are working on including docstrings and corresponding documentation. The site is located here: https://gee-community.github.io/geeViz/build/html/index.html
-
Map Layer area charting option - Can now include any map layer for area summarizing using the
"canAreaChart" : Trueinside thevizparameters dictionary. ThelcmsViewerExample.pyand newmapBiomasViewerExample.pyinclude the area charting functionality. -
Map Biomas Example - There is an example of how to visualize and explore the Map Biomas land use / land cover datasets provided in
mapBiomasViewerExample.py. -
Linearly Interpolated Color Ramp - A function to linearly interpolate a set of hex colors given a starting set of colors, and min and max values (
geeview.get_poly_gradient_ct). This is useful for charting values.
- When querying layers, if
"autoViz" :Trueor aqueryDictis provided, but there is a queried value that is returned not found in that dictionary, the value will be shown instead. Previously, this would result in an error.
- Map Layer Drag Reordering Deactivation Option - Can now deactivate layer reordering using
Map.setCanReorderLayers(False).
- Layer reordering with > 10 layers bug fixed. Sorting > 10 would sort 1,10,11,2,3,....etc.
geeVectorlayerType is now frozen and cannot be reordered. This is because Google Maps API doesn't treat vector layers the same as raster overlays.Support for reordering vectors on the map is not supported.
-
Map Layer Drag Reordering - Can now reorder non-timelapse map layers by clicking and dragging the layer up or down. This feature still can be buggy depending on the underlying layer type.
-
CCDC output smart join method - The
batchFeatherCCDCImgswill take two time series of CCDC coefficients (typically created usingpredictCCDC), and linearly weight which coefficients are being used across a user-specified transition period (featherStartYr,featherEndYr). This allows for a smooth way of splicing together two sets of CCDC outputs with a generous overlapping time period. An example script of how to use this will be provided in a future release once this function solidifies a bit.
- The reducer in
batchSimpleLTFitwas hard-coded to.maxfor reducing any multi-image inputs for a given band/index. This resulted in nulls or errors in overlapping areas if LandTrendr arrays were being mosaicked. There is now a parametermosaicReducerthat can be set to handle overlapping values as you see fit. For array formatted outputs, it has to beee.Reducer.firstNonNull()oree.Reducer.lastNonNull()(default) since GEE doesn't handle array reductions well for imageCollection reductions.
-
Improved cloudStorage Support - Can now manage cloudStorage assets with the
cloudStorageManagerLib.pymodule. -
Improved
exportToCloudStorageWrapper- TheexportToCloudStorageWrappernow can overwrite running tasks or existing cloudStorage blobs ifoverwrite = Trueis specified. Additionally, support for TFrecords is not handled. -
geemap integration - We are starting to use the
geemappackage when possible. There is a new example scriptgeeViz_and_geemap.pythat we will be building on to illustrate using geemap with geeViz. The first example illustrates how you would take a shapefile, usegeemapto convert it to an ee object, and the use that in ageeVizworkflow.
- If
ee.Initialize(project='someProjectID')is called on prior to importing geeViz, that project will automatically be used
- Easier Colab Availability - Colab links are now provided in each notebook in the examples folder.
- 'addLayer' Viz params
reducerbug fix. When a reducer within avizdictionary is passed to the map with aaddLayercall more than once, geeViz would try to serialize it again resulting in an error. It now accepts it and assumes it's already been serialized.
- Fixed bug when
ee.oauth.get_credentials_path()folder didn't exist already and geeViz tried to store the selected project within it. The folder is now created if it does not exist. This is needed whenee.Authenticatedoes not automatically make the folder.
- Fixed bug where project was not read in if ee was initialized outside of geeViz.
setProjectis run when geeView is imported which sets theproject_idto the project provided inee.Initializeif it was provided.
- Enhanced project support - More robust handling of projects for authentication, as well as geeViz viewer authentication. The same project you are using in Python is now used in
geeView.
geeViewquery of collections with over 5000 image*bands values would not query. Reverted to older getRegion-based query method.
-
Color name support - Colors for
vizParamspaletteandclassLegendDictcan now be provided as standard w3 color name strings. -
simpleLTFit batchSimpleLTFit multBy support - Better support for ingesting landTrendr array vertex outputs that are multiplied by 10000 using the
multByparameter in thesimpleLTFitbatchSimpleLTFitfunctions. E.g. if the fitted vertex values were multiplied by 10000, setmultBy = 0.0001.
-
Simplified simpleLANDTRENDR - the
simpleLANDTRENDRfunction has been reworked to be more streamlined, but still provide the same functionality. Steps it uses are now available as stand-alone functions. These include the following new functions:runLANDTRENDR, multLT, LTLossGainExportPrep, and addLossGainToMapand the following previously existing functions:simpleLTFit and convertToLossGain. TheLANDTRENDRViz.py,LANDTRENDRWrapper.py, andLANDTRENDRWrapperNotebook.ipynbexamples have all been updated to utilize these reworked functions. -
setQueryPrecision for Charting - The precision of query outputs is now handled better. It can be changed by using the
Map.setQueryPrecisionfunction. Any floating point number will be constrained by the maximum ofchartPrecisionorchartDecimalProportion*len(someFloatingPointNumber). The default is3and0.25respectively. E.g. if the number is0.12345,max[3,ceiling(len(0.12345)*0.25)] = 3, so the final number would be0.123. -
setQueryDateFormat for Charting - The date format can be changed by using the
Map.setQueryDateFormatfunction orqueryDateFormatproperty within the viz params for aMap.addLayerorMap.addTimeLapsecall. E.g. if you want to only show the year in a chart, you'd putMap.setQueryDateFormat('YYYY')or if you need hours and minutes,Map.setQueryDateFormat('YYYY-MM-dd HH:mm')
- ImageCollection query bug when some pixels were null, but there was a
queryDictprovided is now fixed
-
cloudScore+ For Sentinel-2 - cloudScore+ is now available for cloud and cloud shadow masking for Sentinel-2. It is available for all Sentinel-2 functions/wrappers/examples.
-
mosiac time lapses - if an input imageCollection for an
addTimeLapsecall has multiple images per date, you can now have them mosaicked on-the-fly by using{'mosaic':True}in the viz parameters. -
Reducers for imageCollections in addLayer - When adding an imageCollection with
Map.addLayer, you can now specify the reducer that is used to reduce the imageCollection to a single image to show on the map. Use{'reducer':ee.SomeReducer()}in the viz params.
- Improved Y axis label overcrowding image query charting robustness
- queryDict Chart Y Tick Labels - Charting any imageCollection from a
Map.addLayerorMap.addTimeLapsecall will automatically use thequeryDictvizparameter (e.g.Map.addLayer(someLayer,{"queryDict":{1:"Trees",2:"Grass",3:"Water"}},"SomeLayerName")) to label the Y axis ticks with class names. If class names are too long, they will be shortened. The max character length and the max characters per line in a Y axis tick label can be changed usingMap.setYLabelMaxLengthandMap.setYLabelBreakLengthrespectively.
- Improved array and mixed array and traditional image query charting robustness
- extractPointValuesToDataFrame -
gee2Pandas.extractPointValuesToDataFramefunction extract values as a Pandas data frame. Will handle images or imageCollections automatically.
- Minor bug fixes
- imageArrayPixelToDataFrame -
gee2Pandas.imageArrayPixelToDataFramefunction to easily visualize image array values as a Pandas data frame. - new_interp_date -
changeDetectionLib.new_interp_dateexperimental function to interpolate dates. This method is likely no faster than the previous method (changeDetectionLib.linearInterp), but does extrapolate in a more expected manner.
- Minor bug fixes
-
setZoom -
setZoomfunction to set the zoom level of the map withingeeView.Mapper. This functionality is also avail able in thecenterObjectfunction where the zoom level can optionally be set. -
Date Interpolation now optional for annualizing CCDC -
annualizeCCDCandgetTimeImageCollectionFromCompositesfunctions now can have the linear interpolation turned off. Turning interpolation off will speed up creating outputs, but will result in null values where any date image data are missing.
- Minor bug fixes
- Minor bug fixes
- Minor bug fixes
- robust_featureCollection_to_df -
robust_featureCollection_to_dfin thegee2Pandasmodule will handle large featureCollection conversion to a Pandas dataframe.
create_assetwill now handle nested folder creation withrecursive = True.
- verbose parameter -
getImagesLibLandsat and Sentinel 2 functions now can be less verbose when run. This helps clean up the console. By defult,verbose = False.
assetManagerLibhas improved error handling when various operations cannot be performed - likely due to permissions errors.
- geeView Vertex AI Workbench lab Support -
geeViewcan now be used in Vertex AI notebooks. This support may present authentication bugs depending on how you are authenticating to GEE. You must provide the URL from which the lab is being run as the variableproxy_url. This can be set asMapObject.proxy_url = https://code-dot-region.notebooks.googleusercontent.com/. If it is not provided before a.view()call is made, the user will be prompted to enter it. This attribute will stick with the Map object for the duration of its existence.
- exportToAssetWrapper overwrite support in wrappers -
getLandsatWrapper,getSentinel2Wrapper, andgetLandsatAndSentinel2HybridWrappercan now be run withoverwriteset toTrueorFalse. It will check to see if the asset either exists or is currently being exported. If set toTrueit will stop the export or delete the existing asset and restart it. If set toFalse, it will not start the export.
- exportToAssetWrapper overwrite support -
exportToAssetWrappercan now be run withoverwriteset toTrueorFalse. It will check to see if the asset either exists or is currently being exported. If set toTrueit will stop the export or delete the existing asset and restart it. If set toFalse, it will not start the export.
- geeView Colab Support -
geeViewcan now be used in Google Colab notebooks. This support may present authentication bugs depending on how you are authenticating to GEE.
- More robust authentication and initialization - There are many inconsistencies being introduced for authenticating and initializing to GEE. A new function (
geeViz.geeView.robustInitializer) attempts to handle some of these scenarios. It is not solid though and if you know your particular environment setup, it is best to authenticate and initialize before importing geeViz. There is also difficulty with authenticating on the javaScript client for geeView for some GEE accounts. If the javaScript instance fails to initialize, it will fall back on an existing auth proxy. Since this uses an account different from your own, it may result in errors in accessing assets for viewing in geeView. This can be solved by sharing assets publically.
changeDetectionLib.getTimeImageCollectionFromCompositeshas been updated to fill in a blank image for any missing years and allow a year range to be specified to allow for interpolation and extrapolation. The unused parameters ofstartJulianandendJulianare no longer used.
-
Folium based simple GEE object viewer - A Folium-based GEE object viewer is now available (
foliumView.py). The syntax is very similar to geeView. It tends to load faster, but be quite buggy with layer ordering, and lacks the ability to query values of layers. Examples are provided to help use it (examples\foliumViewer.ipynb,examples\geeViewVSFoliumViewerExampleNotebook.ipynb) -
GEE2Pandas data science helper module - A new module geared toward going between traditional data formats (csv, Excel, dbf, json, etc) and GEE (
gee2Pandas.py). Functions are provided to go from a Pandas dataframe to GEE featureCollection and back. An example is provided (examples\gee2PandasExample.ipynb)
- geeView layer names with odd characters are now accepted (
e.g. / \)
-
Default Query Output Location Change - Query outputs are not placed in the side pane where the legend is located.
- To back to using the default on-map infoWindow use:
Map.setQueryToInfoWindow()
- To back to using the default on-map infoWindow use:
-
Improved Upload To Asset Capabilities - Can now upload tifs to gee assets more easily and handle setting a number of paramters (
assetManagerLib.uploadToGEEImageAsset) and (assetManagerLib.ingestImageFromGCS)
- All MODIS collections have been updated to newer 061 collections
- Query box size now reflects the chosen scale
- Query (inspector) parameters setting - Can now set the click query projection (crs, scale and/or transform), and query box color parameters.
- To set the query box color:
Map.setQueryBoxColor(hexColor) - To set the query crs:
Map.setQueryCRS(crs) - To set the query transform:
Map.setQueryTransform(transform)(note: this will set the scale to null) - To set the query scale:
Map.setQueryScale(scale)(note: this will set the transform to null)
- To set the query box color:
- LandTrendr Array Image Support - Can now export the raw LandTrendr array image output for vertex values only as well as RMSE (
changeDetectionLib.rawLTToVertices). Functions to annualize vertex-only array images are now available too (changeDetectionLib.simpleLTFitwitharrayMode = True. See the LANDTRENDRWrapper example script for a detailed example. - Improved
Map.addLayererror handling - Any ee object added using aMap.addLayercall that fails to load will show an error and not load onto the map rather than stopping the entire map from loading.
- Landsat and Sentinel 2 resampling was set for all bands, including the qa bits. This resulted in speckling around the edges of cloud, cloud shadow, and snow masks. Now, the resampling method remains nearest neighbor for any qa bit band for both Landsat and Sentinel 2. This bug does not exist for MODIS since all masking is performed on continuous bands or the underlying mask of the thermal data.
- Join Collections Using Different Field Names - The
joinFeatureCollectionsandjoinCollectionsfunctions now optionally support different field names between the two collections. e.g.joinFeatureCollections(primaryFC,secondaryFC,'primaryFieldName','secondaryFieldName')
- Sentinel 1 Basic Processing - Integration of Guido Lemoine's shared GEE adaptation of the Refined Lee speckle filter from this script as coded in the SNAP 3.0 S1 Toolbox. This can be found in the
getS1function in thegetImagesLib.
- When loading very large/complicated outputs into geeView, it would often not load the first time. This bug has been fixed to the page properly loads even with very large/complicated EE objects. It can take a while to load the page however.
- geeView Service Account Options - geeView can now utilize a service account key for authentication to GEE. For general guidance for setting up a service account, see this. This service account must be white-listed using this tool. Be sure to download the json key to a local, unshared location. To have geeView use a service account key, specify the path to the json key file as the
serviceKeyPathattribute of the Map object (e.g.Map.serviceKeyPath = r"c:\someFolder\keyFile.json"). This will cause geeView to use that file to gain access to GEE instead of using the default persistent refresh token. If it fails, geeView will then try to use the persistent credential method.
- geeView Token Options - geeView now tries to utilize the default location GEE refresh token instead of a proxy. It cannot use private keys from service accounts. You can however specify a different location of the refresh token from the default (e.g.
Map.accessToken = r'C:\someOtherFolder\someOtherPersistentCredentials). If geeView fails to find a refresh token, it will fall back to utilizing the default proxy location. If that fails, a message will appear. It is best to simply ensure there is a working refresh token available (most easily created usingearthengine authenticateoree.Authenticate()).
- Landsat 9 Integration - Landsat 9 is now included for all Collection 2 Landsat functions.
- Specify geeView port - You can now specify which port to run geeView through by specifying it (e.g.
Map.port = 8000).
- GFS Time Lapse Example - A new example script illustrating how to visualize the GFS forecast model outputs.
- When running geeViz from inside certain IDEs (such as IDLE), the use of sys.executable to identify how to run the local web server would hit on the pythonw instead of python executable. If it finds a pythonw under the sys.executable variable, it now forces it to use the python (without a w).
-
Improved LandTrendr decompression method. - A new function called batchSimpleLTFit has been added to the changeDetectionLib to help provide a faster method for decompressing the LandTrendr stacked output format into a usable time series image collection with all relevant metrics from LandTrendr.
-
LANDTRENDRViz example script. - A new example script that demonstrates how to visualize and post-process LandTrendr outputs.
-
Common projection info dictionary - In order to help organize common projections, a common_projections dictionary is now provided in the getImagesLib module.
- Landsat Collection 2 Level 2 data often have null values in the thermal band. Past versions of geeViz forced all bands to have data (some earlier scenes would have missing data in some but not all bands). This would result in null values in all bands over any area the new Collection 2 surface temperature algorithm could not compute an output. In order to handle this, for Landsat data, all optical bands found in TM, ETM+, and OLI sensors (blue, green, red, nir, swir1, swir2) must have data now, thus allowing a null thermal value to carry through.
- Related to the bug fix above - The medoidMosaicMSD function would still result in a null output if any band had null values. In order to fix this, the min reducer was swapped with the qualityMosaic function and the sum of the squared differences is multiplied by -1 so the qualityMosaic function can function properly. This results in almost the identical result. As the sum of the squared differences approaches 0 for more than a single observation for a given pixel, this method can result in a slightly different pixel choice, but should not make any substantive differences in the final composite.
- Removed the layerType key from the vizParams found within the getImagesLib. If imageCollections were added to the map using addLayer or addTimeLapse, they would not render if using any of the vizParams (vizParamsFalse and vizParamsTrue) since the layer type was explicitly specified as geeImage. You can add the relevant property back into those dictionaries in order to speed up map rendering. Example scripts that make use of these vizParams dictionaries have been updated. e.g. getImagesLib.vizParamsFalse['layerType']= 'geeImage' or getImagesLib.vizParamsFalse['layerType']= 'geeImageCollection'
-
LCMAP_and_LCMS_Viewer Script and Notebook. - A new example viewer that displays two US-wide change mapping products - Land Change Monitoring, Assessment, and Projection (LCMAP) produced by USGS and the Landscape Change Monitoring System (LCMS) produced by USFS. Land cover, land use, and change outputs from each product suite are displayed for easy comparison. The notebook facilitates the comparison by bringing in each set of data from the data suites into separate viewers.
-
Task Tracking in Example Scripts - While the taskManagerLib is not new, the task tracking functionality available within that module was added to the example scripts that export data.
- Fixed bug in pulling CCDC most recent loss and gain year out. It now behaves in a similar manner as the most recent loss and gain outputs.
- Time Lapse Charting - Any time lapse that is visible will be charted if double-clicked using "Query Visible Map Layers" tool.
- Landsat Collection 2 Support - Landsat Collection 2 is now used by default for all getLandsat methods. Collection 1 is still available by specifying landsatCollectionVersion = 'C1'. Specify 'C2' (which is the default) if you would like to use Collection 2.
- None
- Fixed endDate for all filterDate calls. Previously, it had been assumed this was inclusive of the day of the endDate, when the GEE filterDate method is exclusive of the day of the endDate. Since the assumption is that all dates specified are inclusive, the current fix involves advancing all endDates provided to any filterDate function by 1 day.
- LANDTRENDRWrapper Notebook - A new example of how to use LandTrendr in a Jupyter Notebook format. This is very similar to the script, but provides a little more detail of the resources available to run LandTrendr.
- geeView would not work in ArcPro Miniconda Python builds (and perhaps others). The new fix addresses this issue and allows it to run from Miniconda Python builds based from ArcPro
- Smarter GEE initialization - All modules that call upon ee.Initialize will now check to ensure GEE hasn't already been initialized. If it hasn't, it will initialize.
- GEE occasionally wouldn't recognize an imageCollection as such when adding a TimeLapse. An explicit cast to imageCollection was now made.
- Back and foward view buttons - Users can go backward and forward views within the geeView viewer.
- Fixed export wrappers exporting empty areas with GEE update to export methods
- Notebook examples - Several examples are now available in an interactive notebook format under the examples module.
- CCDC updates - Updates to CCDC annualizing functionality.
- MODIS Processing - Similar pre-processing (cloud and cloud shadow masking) functionality that has been available for Landsat and Sentinel 2 now available for MODIS data.