A command-line interface for Google Maps Engine.

Requires Python 2.7 (32bit).

Note: If you receive a "setuptools pip failed with error code 1" error whilst setting up your virtualenv you need to downgrade to virtualenv 1.11.2 due to this issue.

pip install virtualenv==1.11.2

virtualenv venv
. venv/bin/activate
pip install --editable .

If you don't already have Python installed check out ActivePython.

virtualenv venv
venv\Scripts\activate.bat
pip install --editable .

Note: If you receive the error "error: Unable to find vcvarsall.bat" whilst Hodor is installing you need to install a C compiler due to this issue. Simply download Visual Studio C++ 2008 Express Edition, open another command prompt, and away you go.

The first time you run Hodor he will open a browser window and prompt you to authorise him to access Google Maps Engine on your behalf. When the The authentication flow has completed message shows you can close the tab and return to your terminal.

Hodor ships with its own application credentials, so by default you don't need to setup anything else.

However, Hodor's application credentials are limited to 10,000 requests/day, so it is possible for other users of Hodor to exhaust this shared pool.

If you find this occurring (a) let me know, and (b) you can setup your own applications credentials thusly:

1. Go to the Google Developers Console and Create a new project
2. Go to the APIs & Auth section and enable the Google Maps Engine API
3. Click on Credentials and create a new Installed Application
4. Choose Download JSON and copy it as oauth.json to the directory you have Hodor installed to.
5. Remove the existing credentials-store.json file.

Hodor knows about:

• Uploading and creating raster and vector assets en masse, and adding or creating layers and raster collections from them.
• Querying vector tables to retrieve features (multi-threaded).
• Faux Table Replace by which an existing vector table can have its data replaced with a new set of files.
• Modifying the contents of vector tables (WIP).
• Updating record-level vector data (WIP).
• A range of random tools for doing things with GME.

To work out what capabilities a certain Hodor command has simply pass --help on the end of the command.

To begin using Hodor, first activate your virtual environment:

Linux

. venv/bin/activate

Windows

venv/Scripts/activate.bat

Hodor can upload raster and vector data to create new assets and, optionally, then directly create new layers or append them to an existing raster mosaic or layer.

You can pass Hodor three types of configstore:

1. A config.json file representing a single asset.
2. A config.json file representing multiple assets (bulk loading).
3. A path containing multiple directories, each with config.json files, representing single or multiple assets.

Hodor's recommended directory structure:

org-name/        -- Name of the your organisation (e.g. Landgate)
  asset-name-1/       -- Name of the dataset (e.g. lgate_cadastre_poly_1)
    config.json     -- Store your asset metadata here
    layers.json     -- Store layer metadata here (optional)
    payload/        -- Store your data here (e.g. Shapefiles, JPEG2000)
  asset-name-2/
    ...

For example:

Upload one or more raster assets:

hodor create raster "data/org-name/asset-name-1/config.json"

Upload multiple vector assets:

hodor create vector "data/org-name/"

Upload a new raster and add it to an existing raster mosaic:

hodor create raster --mosaic-id={asset-id} "data/org-name/asset-name-1/config.json"

Hodor supports the bulk ingestion of raster and vector based on a single template config.json file.

Hodor will create one new asset for every unique set of files in your payload directory. Each asset will be based on the template information provided in config.json, with the name the asset being derived from the name of the file.

For example:

payload/
  file1.tif
  file1.tfw
  file2.tif
  file2.tfw

Will create two assets named file1 and file2, each with a TIF and TFW file associated with it.

When creating a new asset (currently only vector data) Hodor will automatically creating any layers specified in a layers.json file that resides in the same directory as your config.json file.

Hodor understand raster mosaics and layers, so when you create new raster assets Hodor can:

• Add the newly created raster(s) to the mosaic supplied by --mosaic-id. Optionally, you may also trigger the mosaic to begin processing with --process-mosaic.
• Add the newly created raster(s) to the layer supplied by --layer-id.

config.json should contain your asset metadata fields in JSON format as defined by Google Maps Engine.

See the GME API documentation for the minimum required fields for vector and raster data.

{
  "projectId": "{your-gme-project-id}",
  "name": "Alkimos 1964 Hodor",
  "draftAccessList": "Map Editors",
  "attribution": "Landgate",
  "rasterType": "image",
  "description":"This is a description.",
  "tags" : ["hodor"],
  "maskType":"imageMask",
  "acquisitionTime": {
    "start" : "2010-01-24T00:00:00Z",
    "end" : "2010-01-25T00:00:00Z",
    "precision" : "day"
  }
}

layers.json should contain your layer metadata fields in JSON format as defined by Google Maps Engine.

See the layer create documentation for the minium required fields.

{
  "projectId": "06151154151057343427",
  "layers": [
    {
      "name": "daa_003_subset layer",
      "datasourceType": "table",
      "draftAccessList": "Map Editors",
      "tags": ["Hodor"],
      "styleFile": "style.json",
      "infoWindowFile": "infoWindow.html"
    }
  ]
}

layers.json supports two optional fields:

styleFile is an optional parameter that will use the given file as the layer's style document.

infoWindowFile is an optional parameter that will read the given HTML file into the layer's style document.

Hodor can query vector tables and export the results to the GeoJSON format, including:

• Limiting the query to a specific bounding box and GME's SQL-like query language.
• Querying huge areas by splitting the bounding box into manageable chunks.
• Multi-threaded querying to send many requests to GME in parallel for optimised retrieval.

Return all of the features from a vector table:

hodor features list {table-id} features.json

Return all of the features from a vector table in a given area and with a LOCALITY of NANNUP:

hodor features list -bbox="115.876581, -31.926812, 115.960266, -31.893442" -where="LOCALITY <> 'NANNUP'" {table-id} features.json

Some associated documentation for anyone wishing to build on or understand this wee tool.

Installing the Google Maps Engine API Client Library for Python

The PyDoc reference for the Google Maps Engine API

APIs Explorer for the Google Maps Engine API

Google API Discovery Service for GME

Due credit and thanks go to the folks that put together the Google Cloud Platform code samples, particularly the chunked transfer example on which Hodor was originally based.

And to Armin Ronacher for his fantastic Click library.