Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Tests for all combinations of settings #20

Open
svdhoog opened this issue May 26, 2018 · 2 comments
Open

Tests for all combinations of settings #20

svdhoog opened this issue May 26, 2018 · 2 comments

Comments

@svdhoog
Copy link
Owner

svdhoog commented May 26, 2018

Test data sets

See: circles model.

Config.yaml settings

  • Types of analysis (4)
    agent, multiple_run, multiple_batch, multiple_set

  • Types of plots (4)
    timeseries, boxplot, histogram, scatterplot

  • Types of statistical summary (9)
    full, mean, median, quantile, custom_quantile, maximum, minimum, upper_quartile, lower_quartile

  • Nature of Plots (2) (option is set per analysis, ie per variable or entry in the config.yaml file)
    one, many
    one: a single plot with all curves or plots is produced in 1 file
    many: multiple plots are produced in multiple files, with 1 file per curve or plot

Total combinations: 4x4x9x2 = 288

Some in-build exceptions that have been explicitly excluded:

  • agent_analysis + one_output : "Quantile not possible for agent level analysis!"
    The reason is that an agent only has scalar values, so there is no range for quantiles.

  • agent_analysis + many_output : "- Warning: too many plots will be produced !!! "
    This is a precautionary warning.

  • agent_analysis + custom_quantiles : "Quantile not possible for agent level analysis"
    The reason is that an agent only has scalar values, so there is no range for quantiles.

  • scatter_plot + custom_quantiles: "Delay not supported for Quantiles! Adjust parameter and retry!"
    This combinations has not yet been implemented.

General notes

  • summary==full + many_output : Does not make sense? Since summary==full creates the full ensemble of data, so there is only one data set to plot. Hence, this will always result in 1 plot in 1 file. So here many_output has no effect, and always yield the same result as one_output.
@svdhoog
Copy link
Owner Author

svdhoog commented Jun 13, 2019

Test Circle model

The test data allows to test these settings:

  • Types of analysis (4)
    agent, multiple_run, multiple_batch, multiple_set

  • Types of plots (4)
    timeseries, boxplot, histogram, scatterplot

  • Types of statistical summary (3)
    full, mean,custom_quantile

[the remaining 6 options are: median, custom_quantile, maximum, minimum, upper_quartile, lower_quartile]

  • Nature of Plots (2)
    one, many

Total combinations: 4x4x3x2 = 96

@svdhoog
Copy link
Owner Author

svdhoog commented Jun 13, 2019

Scatterplot mean

one plot

summary!=full

plot 61

Shows a summary across agents, for each run, each set (hence: 8 circles)

FLAViz-scatter-oneplot-summarymean-61-63

Notes:

  • The first circle (set=1, run=1) is centred on (a,b)=(3.5,1) because we have 6 agents, with a=1...6, b=1. The means are E[a]=3.5 and E[b]=1, resp.
  • The 2nd circle (set=1, run=2) is centred on (a,b)=(3.5,2) because b=run_no.
  • All concentric circles that are centred on (a,b)=(3.5,1) belong to run 1 and increase with the set number.
  • All concentric circles that are centred on (a,b)=(3.5,2) belong to run 2 and increase with the set number.

plot 62

Shows a summary across agents & runs, for each set (hence: 4 circles)

Notes:

  • The circle for set 1 is centred on (a,b)=(3.5,1.5) because we are now averaging across runs, and the two blue circles in plot60 were centred on b=1 and b=2, resp.
  • The circles for sets 2-4are again increasing in radius with the set number.

plot 63

shows summary across agents & runs & sets (hence: 1 circle)

Notes:

  • The single circle is the average across the 4 circles in plot62, which were centred on a=3.5 with b=1...4. Hence the average is centred on a=3.5,b=2 with an average radius of r=2.

many plot

FLAViz-scatter-manyplot-summarymean-65-67

Plot 64

Shows agent-level analysis with many plots. This produces 1 file per (set,run,agent) combination (48 files for the 48 circle plots in plot 60).

Plot 65

Shows multi-run analysis with many plots. This produces 1 file per (set,run) combination (8 files for the 8 circle plots in plot 61).

Plot 66

Shows multi-batch analysis with many plots. This produces 1 file per set (4 files for the 4 circle plot in plot 62).

Plot 67

Shows multi-set analysis with many plots. This produces 1 file (1 files for the 1 circle plot in plot 63).

Scatterplot full

one plot

summary=full

FLAViz-scatter-oneplot-summaryfull-71-73

Plot 70

Summary=full is not possible with agent-level analysis.

Plot 71

Plot 72

Plot 73

many plot

summary=full

FLAViz-scatter-manyplot-summaryfull-75-77

Plot 74

Summary=full is not possible with agent-level analysis.

Plot 75

Plot 76

Plot 77

svdhoog added a commit that referenced this issue Jun 13, 2019
svdhoog added a commit that referenced this issue Jun 13, 2019
Issue #20: Test cases for Scatterplots plot60
svdhoog added a commit that referenced this issue Jun 14, 2019
Issue #20: Test cases for Scatterplots; final fixes filenames
svdhoog added a commit that referenced this issue Jun 14, 2019
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant