Python tSNE failed about algorithm-repository

[bldrvnlw]

Some bug in the charting logic by the looks of it, I'll look into it.

[bldrvnlw]

I've examined the testing data and the log above in detail notice a couple of things which are leading to problems.

1. There are three possible table which could be being used here: cde_feature_a, cde_feature_b, cde_feature_c. Of these only the _b and _c table contain data in the minimentalstate column, but in cde_features_a all the values are NULL so the PARAM_query won't return anything .

Early on there is a "mean of empty slice" error I assume is related to getting empty data from the select.

2. No covariables are specified for the tsne dimensions.

In the python-tsne/test/docker-compose.yml there is a good example of how to pass the parameters (in this case the dependent variable is name and this is also used to label output in the graph). Note the multiple covariables (4 in total) goving a 4 dimensional input.:

  PARAM_variables: "name"
  PARAM_covariables: "sepal_length,sepal_width,petal_length,petal_width"
  PARAM_grouping: ""
  PARAM_query: "SELECT name,sepal_length,sepal_width,petal_length,petal_width FROM iris"
  PARAM_meta: "{\"name\":{\"code\":\"name\",\"type\":\"string\"},\"sepal_length\":{\"code\":\"sepal_length\",\"type\":\"real\"},\"sepal_width\":{\"code\":\"sepal_width\",\"type\":\"real\"},\"petal_length\":{\"code\":\"petal_length\",\"type\":\"real\"}, \"petal_width\":{\"code\":\"petal_width\",\"type\":\"real\"}}"
  PARAM_MODEL_dependent_is_label: "True"
  PARAM_MODEL_perplexity: 30
  PARAM_MODEL_do_zscore: "True"
  PARAM_MODEL_iterations: 1000
  PARAM_MODEL_theta: 0.5
  PARAM_MODEL_target_dimensions: 2

[bldrvnlw]

Given the comments about I think this is mainly a usage problem. Based on that I'm closing the issue but feel free to re-open it as a documentation issue (or open a new issue) and I can provide better guidance in the Readme.md file for using the python/tsne