Add api fucntionality

datasetter/api.py

@@ -0,0 +1,78 @@
+import forge
+from typing import Optional
+import pandas as pd
+import json
+
+
+def as_json(doc):
+    return json.loads(
+        pd.Series([doc])
+        .to_json(orient='records', date_format='iso')
+        )[0]
+
+
+def add_dataset(dataset, fast_api, uri):
+    """Create FastAPI endpoints to serve the dataset.
+
+    Parameters
+    ----------
+    fast_api : fastapi.applications.FastAPI
+        A FastAPI application object. A new endpoint will be added in this application.
+    uri : str
+        The relative uri where the endpoint will be added.
+    dataset : datasetter.dataset.Dataset
+        The dataset object to be served.
+
+    Returns
+    -------
+    Nothing : The endpoint is created as a side effect (inplace) in the `fast_api` application.
+    """
+    uri = '/' + uri.strip('/')
+    facet = forge.kwarg('facet', type=str)
+    rows = forge.kwarg('rows', default=10, type=Optional[int])
+    skip = forge.kwarg('skip', default=0, type=Optional[int])
+    kwargs = [forge.kwarg(facet, default=None, type=Optional[str])
+              for facet in dataset.facets]
+
+    @fast_api.get(uri + "/")
+    def get_metadata():
+        return as_json(dataset.metadata)
+
+    @fast_api.get(uri + "/count")
+    @forge.sign(*kwargs)
+    def count(**kwargs):
+        filters = {key: val for key, val in kwargs.items() if val is not None}
+        count = dataset.count(**filters)
+        return as_json({
+            "count": int(count),
+            "filters": filters,
+            })
+
+    @fast_api.get(uri + "/count-by/{facet}")
+    @forge.sign(facet, rows, skip, *kwargs)
+    def count_by(facet, rows=10, skip=0, **kwargs):
+        filters = {key: val for key, val in kwargs.items() if val is not None}
+        result = dataset.count_by(facet, rows=rows, skip=skip, **filters)
+        return as_json({
+            "facet": facet,
+            # "count": len(result),  # TODO : add "nunique" feature in count_by schema
+            "rows": len(result),
+            "skip": skip,
+            "filters": filters,
+            "data": {str(key): int(val) for key, val in result.items()},
+            })
+
+    @fast_api.get(uri + "/sample")
+    @forge.sign(rows, skip, *kwargs)
+    def sample(rows=10, skip=0, **kwargs):
+        filters = {key: val for key, val in kwargs.items() if val is not None}
+        result = dataset.sample(rows=rows, skip=skip, **filters)
+        count = dataset.count(**filters)
+        return as_json({
+            # "facet": facet,
+            "count": count,
+            "rows": len(result),
+            "skip": skip,
+            "filters": filters,
+            "data": result.to_dict(orient='records'),
+            })

datasetter/dataset.py

@@ -1,17 +1,4 @@
 # Dataset : base class for datasetter
-
-from typing import Optional
-import pandas as pd
-import json
-
-
-def as_json(doc):
-    return json.loads(
-        pd.Series([doc])
-        .to_json(orient='records', date_format='iso')
-        )[0]
-
-
 class FacetUnavailableError(Exception):
     """This class is used to raise exceptions due to unavailable facets."""
     pass
@@ -78,68 +65,3 @@ def sample(self, rows=10, skip=0, **filters):
             A sample from the data.
         """
         raise NotImplementedError()
-
-    def fastapi_serve(self, fast_api, uri):
-        """Create FastAPI endpoints to serve the dataset.
-
-        Parameters
-        ----------
-        fast_api : fastapi.applications.FastAPI
-            A FastAPI application object. A new endpoint will be added in this application.
-        uri : str
-            The relative uri where the endpoint will be added.
-
-        Returns
-        -------
-        Nothing : The endpoint is created as a side effect (inplace) in the `fast_api` application.
-        """
-        uri = '/' + uri.strip('/')
-        import forge
-        facet = forge.kwarg('facet', type=str)
-        rows = forge.kwarg('rows', default=10, type=Optional[int])
-        skip = forge.kwarg('skip', default=0, type=Optional[int])
-        kwargs = [forge.kwarg(facet, default=None, type=Optional[str])
-                  for facet in self.facets]
-
-        @fast_api.get(uri + "/")
-        def get_metadata():
-            return as_json(self.metadata)
-
-        @fast_api.get(uri + "/count")
-        @forge.sign(*kwargs)
-        def count(**kwargs):
-            filters = {key: val for key, val in kwargs.items() if val is not None}
-            count = self.count(**filters)
-            return as_json({
-                "count": int(count),
-                "filters": filters,
-                })
-
-        @fast_api.get(uri + "/count-by/{facet}")
-        @forge.sign(facet, rows, skip, *kwargs)
-        def count_by(facet, rows=10, skip=0, **kwargs):
-            filters = {key: val for key, val in kwargs.items() if val is not None}
-            result = self.count_by(facet, rows=rows, skip=skip, **filters)
-            return as_json({
-                "facet": facet,
-                # "count": len(result),  # TODO : add "nunique" feature in count_by schema
-                "rows": len(result),
-                "skip": skip,
-                "filters": filters,
-                "data": {str(key): int(val) for key, val in result.items()},
-                })
-
-        @fast_api.get(uri + "/sample")
-        @forge.sign(rows, skip, *kwargs)
-        def sample(rows=10, skip=0, **kwargs):
-            filters = {key: val for key, val in kwargs.items() if val is not None}
-            result = self.sample(rows=rows, skip=skip, **filters)
-            count = self.count(**filters)
-            return as_json({
-                # "facet": facet,
-                "count": count,
-                "rows": len(result),
-                "skip": skip,
-                "filters": filters,
-                "data": result.to_dict(orient='records'),
-                })

test/test_api.py

@@ -0,0 +1,13 @@
+# import pytest
+import numpy as np
+import pandas as pd
+
+from datasetter.api import as_json
+
+
+def test_as_json():
+    assert as_json(None) is None
+    assert as_json(pd.NaT) is None
+    assert as_json(np.NaN) is None
+    assert as_json({}) == {}
+    assert isinstance(as_json(pd.Timestamp.utcnow()), str)

test/test_dataset.py

@@ -1,16 +1,7 @@
 import pytest
-import numpy as np
 import pandas as pd
 
-from datasetter.dataset import Dataset, as_json
-
-
-def test_as_json():
-    assert as_json(None) is None
-    assert as_json(pd.NaT) is None
-    assert as_json(np.NaN) is None
-    assert as_json({}) == {}
-    assert isinstance(as_json(pd.Timestamp.utcnow()), str)
+from datasetter.dataset import Dataset
 
 
 def test_dataframe():