Instruct file for context

In appletree, we use an instruct json file to initialize Context. Some examples can be found in instructs folder. In the json file, the dictionary must contain three items: “configs”, “par_config” and “likelihoods”.

The value of “configs” is a dictionary that overwrites the default plugins’ config, like

{
    name_of_plugin_config: new_value_of_plugin_config,
    ...
}

We will discuss more about how appletree deals with them in map section.

The value of “par_config” should be the file name of parameter file described in parameter.

The value of “likelihoods” is a dictionary that gives the definition of likelihood. If the fitting is a single dataset fitting, the dictionary only has one item; if the fitting is a combined fitting, it could have multiple items. In each item, it must contain

• “components”: all components that define the model.

• “data_file_name”: data file name, must be .pkl or .csv.

• “bins_type”: can be “meshgrid” or “equiprob”.

• “bins_on”: names of variables that the histogram is applied on. For most of the case, we only recommend “equiprob”, meaning a equi-probable binning defined by data.

• “bins”: number of bins.

• “x_clip”: range of the first dimension.

• “y_clip”: range of the second dimension.

For example,

{
    "components": {
        "rn220_er": "ERBand",
        "rn220_ac": {
            "component_cls": "AC",
            "file_name": "AC_Rn220.pkl"
        }
    },
    "copy_parameters": {
        "rn220_er_rate": "er_rate",
        "rn220_ac_rate": "ac_rate"
    },

    "data_file_name": "data_Rn220.csv",
    "bins_type": "equiprob",
    "bins_on": ["cs1", "cs2"],
    "bins": [15, 15],
    "x_clip": [0, 100],
    "y_clip": [2e2, 1e4]
}

In this likelihood config, the data has two components: ER and AC, where ER component is ERBand, and AC component is AC with input file “AC_Rn220.pkl”. And the binned Poisson likelihood is defined in the equi-probable binning of cS1-cS2 space, with 15 bins on each dimension. To understand how appletree locates the files and classes, like “data_Rn220.csv” etc., check file section.