Add basic examples

By Erik-Jan van Kesteren

examples/basic_example.json

@@ -0,0 +1,90 @@
+{
+    "n_rows": 5,
+    "n_columns": 5,
+    "provenance": {
+        "created by": {
+            "name": "MetaSynth",
+            "version": "0.1.2+15.ged3af36",
+            "privacy": null
+        },
+        "creation time": "2022-11-17T13:54:16.686166"
+    },
+    "vars": [
+        {
+            "name": "ID",
+            "type": "discrete",
+            "dtype": "<class 'polars.datatypes.Int64'>",
+            "prop_missing": 0.0,
+            "distribution": {
+                "name": "UniqueKeyDistribution",
+                "parameters": {
+                    "low": 1,
+                    "consecutive": 1
+                }
+            }
+        },
+        {
+            "name": "fruits",
+            "type": "categorical",
+            "dtype": "<class 'polars.datatypes.Categorical'>",
+            "prop_missing": 0.0,
+            "distribution": {
+                "name": "MultinoulliDistribution",
+                "parameters": {
+                    "labels": [
+                        "apple",
+                        "banana"
+                    ],
+                    "probs": [
+                        0.4,
+                        0.6
+                    ]
+                }
+            }
+        },
+        {
+            "name": "B",
+            "type": "discrete",
+            "dtype": "<class 'polars.datatypes.Int64'>",
+            "prop_missing": 0.0,
+            "distribution": {
+                "name": "PoissonDistribution",
+                "parameters": {
+                    "mu": 3.0
+                }
+            }
+        },
+        {
+            "name": "cars",
+            "type": "categorical",
+            "dtype": "<class 'polars.datatypes.Categorical'>",
+            "prop_missing": 0.0,
+            "distribution": {
+                "name": "MultinoulliDistribution",
+                "parameters": {
+                    "labels": [
+                        "audi",
+                        "beetle"
+                    ],
+                    "probs": [
+                        0.2,
+                        0.8
+                    ]
+                }
+            }
+        },
+        {
+            "name": "optional",
+            "type": "discrete",
+            "dtype": "<class 'polars.datatypes.Int64'>",
+            "prop_missing": 0.2,
+            "distribution": {
+                "name": "DiscreteUniformDistribution",
+                "parameters": {
+                    "low": -30,
+                    "high": 301
+                }
+            }
+        }
+    ]
+}

examples/basic_example.py

@@ -0,0 +1,39 @@
+import polars as pl
+from metasynth import MetaDataset
+
+# example dataframe from polars website
+df = pl.DataFrame(
+    {
+        "ID": [1, 2, 3, 4, 5],
+        "fruits": ["banana", "banana", "apple", "apple", "banana"],
+        "B": [5, 4, 3, 2, 1],
+        "cars": ["beetle", "audi", "beetle", "beetle", "beetle"],
+        "optional": [28, 300, None, 2, -30],
+    }
+)
+
+# convert appropriate columns to categorical
+df = df.with_columns([
+    pl.col("fruits").cast(pl.Categorical),
+    pl.col("cars").cast(pl.Categorical),
+])
+
+# set A to unique and B to not unique
+spec_dict = {
+    "ID": {"unique": True},
+    "B": {"unique": False}
+}
+
+# create metadataset
+mds = MetaDataset.from_dataframe(df, spec=spec_dict)
+
+# write to json
+mds.to_json("examples/basic_example.json")
+
+# then, export json from secure environment
+
+# outside secure environment, load json
+mds_out = MetaDataset.from_json("examples/basic_example.json")
+
+# create a fake dataset
+mds_out.synthesize(10)