Phenofhy

Simulating data locally ​

This tutorial shows how to generate an OFH-like sample dataframe locally using phenofhy.simulate.simulate_phenotype_df().

Use this when you want to test preprocessing and analysis code outside the TRE without querying DNAnexus.

Basic example ​

python
from phenofhy.simulate import simulate_phenotype_df

# Uses default 1000 rows and default 15 OFH-like columns
df = simulate_phenotype_df(seed=42)

df.head()

Choose sample size and fields ​

Pass one entity.field string or a list of fields.

python
from phenofhy.simulate import simulate_phenotype_df

df = simulate_phenotype_df(
    sample=500,
    fields=[
        "participant.pid",
        "participant.birth_year",
        "participant.demog_sex_2_1",
        "participant.demog_ethnicity_1_1",
        "clinic_measurements.height",
        "clinic_measurements.weight",
    ],
    seed=7,
)

Include non-response codes ​

By default, non-response/sentinel codes are excluded. Set include_nonresponse=True if you want them included.

python
from phenofhy.simulate import simulate_phenotype_df

df = simulate_phenotype_df(
    sample=1000,
    fields="participant.demog_ethnicity_1_1",
    include_nonresponse=True,
    seed=123,
)

Control missingness ​

Use a global missing rate or a per-field mapping.

python
from phenofhy.simulate import simulate_phenotype_df

df = simulate_phenotype_df(
    sample=1000,
    fields=[
        "participant.pid",
        "clinic_measurements.height",
        "clinic_measurements.weight",
    ],
    missing_rate={
        "participant.pid": 0.0,
        "clinic_measurements.height": 0.1,
        "clinic_measurements.weight": 0.15,
    },
    seed=2026,
)

Notes ​

  • Valid fields and base types come from data_dictionary.csv.
  • Categorical code domains come from codings.csv.
  • For numeric fields not present in codings, conservative fallback ranges are used.
  • This simulator is intended for local testing and development, not for inference on real OFH participants.