Metadata for Outliers detection algorithm

[larjohn]

The metadata of the algorithm don't contain any input for user defined data - could be done statically.
{
"algorithm": {
"endpoint": [
"DAMUrl",
"/outlier_detection/LOF",
"sample|real"
],
"instance": "outlier_dm.detect_outliers_subpopulation_lattice",
"method": "POST",
"name": "do_outlier_detection_lof",
"prompt": "select one or more files, to detect most isolated items (based on density)",
"title": "outlier Detection based on Local Outlier Factor"
},
"badDataSetPatterns": [
"bad"
],
"badDataSets": [
"baddataset_x",
"baddataset_y"
],
"dataSetPatterns": [
"kilkis"
],
"dataSets": [
"?any",
"kilkis_neu.csv"
],
"decision": true,
"description": "outlier detection based on local outlier factor",
"input": {
"description": "if the endpoint is '/outlier_detection/LOF/sample', no input is needed. The server uses '/Data/Kilkis_neu.csv' as input csv file. Following input is for '/outlier_detection/LOF/real'",
"full_output": {
"cardinality": "1",
"default": "partial",
"guess": false,
"name": "full_output",
"required": false,
"title": "Whether full output",
"type": "string"
},
"output_file": {
"cardinality": "1",
"default": "Result",
"guess": false,
"name": "output",
"required": false,
"title": "Output File Name",
"type": "string"
},
"raw_data": {
"cardinality": "1",
"guess": false,
"name": "filename",
"note": "all input filenames are concatenated by '+', starting with '+', e.g., '+budget-kilkis-expenditure-2012'or '+budget-kilkis-expenditure-2012+budget-kilkis-expenditure-2013' ",
"required": true,
"title": "Select one or more files to identify anormal values",
"type": "turtle"
}
},
"name": "outlierDetection_LOF",
"output": {
"cardinality": "1",
"instance": "json_outout with 'filename' as the key",
"location": "/static/output",
"name": "output",
"type": "CSV"
}
}

[HimmelStein]

please send me following examples.

1. a link of dataset (plain)
2. several links of dataset (aggregate)

we shall update the DAM interface for this algorithm.

[larjohn]

Facts:

http://ws307.math.auth.gr/rudolf/public/api/3/cubes/aragon-2008-income__568a8/facts

Aggregates:

http://ws307.math.auth.gr/rudolf/public/api/3/cubes/global/aggregate?drilldown=global__administrativeClassification__854d0.notation%7Cglobal__administrativeClassification__854d0.prefLabel&order=global__amount__0397f.sum:desc&pagesize=30

[HimmelStein]

I updated DAM, preprocessing_dm, outlier_dm
https://github.com/openbudgets/DAM (staging_indigo branch)
https://github.com/openbudgets/outlier_dm
https://github.com/openbudgets/preprocessing_dm

input json, output json works at my local environments

in the do_outlier_detection_lof function at the app.py file of DAM, please tell me the variable which stores the link, currently, I just use 'filename' for testing.

## 'filename' shall be the variable storing the link from Indigo!!
##
filename = request.args.get('filename', 'http://ws307.math.auth.gr/rudolf/public/api/3/cubes/aragon-2008-income__568a8/facts')