Merge pull request #2850 from finos/python-schema-relax

texodus · web-flow · commit d630f0fd2712 · 2024-11-15T10:36:19.000-05:00
Relax `Table` schema constructor
diff --git a/rust/perspective-python/perspective/__init__.py b/rust/perspective-python/perspective/__init__.py
@@ -16,7 +16,6 @@
     "Server",
     "Client",
     "PerspectiveError",
-    "PerspectiveWidget",
     "ProxySession",
 ]
 
diff --git a/rust/perspective-python/perspective/tests/table/test_table.py b/rust/perspective-python/perspective/tests/table/test_table.py
@@ -348,6 +348,26 @@ def test_table_symmetric_string_schema(self):
 
         assert tbl2.schema() == schema
 
+    def test_table_python_schema(self):
+        data = {
+            "a": int,
+            "b": float,
+            "c": str,
+            "d": bool,
+            "e": date,
+            "f": datetime,
+        }
+
+        tbl = Table(data)
+        assert tbl.schema() == {
+            "a": "integer",
+            "b": "float",
+            "c": "string",
+            "d": "boolean",
+            "e": "date",
+            "f": "datetime",
+        }
+
     # is_valid_filter
 
     # def test_table_is_valid_filter_str(self):
diff --git a/rust/perspective-python/perspective/tests/test_dependencies.py b/rust/perspective-python/perspective/tests/test_dependencies.py
@@ -44,3 +44,10 @@ def test_lazy_modules():
 
     for k, v in cache.items():
         sys.modules[k] = v
+
+
+def test_all():
+    import perspective
+
+    for key in perspective.__all__:
+        assert hasattr(perspective, key)
diff --git a/rust/perspective-python/src/client/client_sync.rs b/rust/perspective-python/src/client/client_sync.rs
@@ -23,6 +23,7 @@ use pyo3::prelude::*;
 use pyo3::types::*;
 
 use super::python::*;
+use crate::py_err::ResultTClientErrorExt;
 use crate::server::PySyncServer;
 
 #[pyclass]
diff --git a/rust/perspective-python/src/client/mod.rs b/rust/perspective-python/src/client/mod.rs
@@ -14,3 +14,5 @@ pub mod client_sync;
 mod pandas;
 mod pyarrow;
 pub mod python;
+pub mod table_data;
+pub mod update_data;
diff --git a/rust/perspective-python/src/client/python.rs b/rust/perspective-python/src/client/python.rs
@@ -10,26 +10,27 @@
 // ┃ of the [Apache License 2.0](https://www.apache.org/licenses/LICENSE-2.0). ┃
 // ┗━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┛
 
-use std::any::Any;
 use std::collections::HashMap;
 use std::str::FromStr;
 use std::sync::Arc;
 
 use async_lock::RwLock;
 use futures::FutureExt;
 use perspective_client::{
-    assert_table_api, assert_view_api, clone, Client, ClientError, OnUpdateMode, OnUpdateOptions,
-    Table, TableData, TableInitOptions, TableReadFormat, UpdateData, UpdateOptions, View,
+    assert_table_api, assert_view_api, clone, Client, OnUpdateMode, OnUpdateOptions, Table,
+    TableData, TableInitOptions, TableReadFormat, UpdateData, UpdateOptions, View,
     ViewOnUpdateResp, ViewWindow,
 };
-use pyo3::create_exception;
 use pyo3::exceptions::PyValueError;
 use pyo3::prelude::*;
-use pyo3::types::{PyAny, PyBytes, PyDict, PyList, PyString};
+use pyo3::types::{PyAny, PyBytes, PyDict, PyString};
 use pythonize::depythonize_bound;
 
 use super::pandas::arrow_to_pandas;
+use super::table_data::TableDataExt;
+use super::update_data::UpdateDataExt;
 use super::{pandas, pyarrow};
+use crate::py_err::{PyPerspectiveError, ResultTClientErrorExt};
 
 #[derive(Clone)]
 pub struct PyClient {
@@ -38,138 +39,6 @@ pub struct PyClient {
     close_cb: Option<Py<PyAny>>,
 }
 
-#[extend::ext]
-pub impl<T> Result<T, ClientError> {
-    fn into_pyerr(self) -> PyResult<T> {
-        match self {
-            Ok(x) => Ok(x),
-            Err(x) => Err(PyPerspectiveError::new_err(format!("{}", x))),
-        }
-    }
-}
-
-create_exception!(
-    perspective,
-    PyPerspectiveError,
-    pyo3::exceptions::PyException
-);
-
-#[extend::ext]
-impl UpdateData {
-    fn from_py_partial(
-        py: Python<'_>,
-        input: &Py<PyAny>,
-        format: Option<TableReadFormat>,
-    ) -> Result<Option<UpdateData>, PyErr> {
-        if let Ok(pybytes) = input.downcast_bound::<PyBytes>(py) {
-            // TODO need to explicitly qualify this b/c bug in
-            // rust-analyzer - should be: just `pybytes.as_bytes()`.
-            let vec = pyo3::prelude::PyBytesMethods::as_bytes(pybytes).to_vec();
-
-            match format {
-                Some(TableReadFormat::Csv) => Ok(Some(UpdateData::Csv(String::from_utf8(vec)?))),
-                Some(TableReadFormat::JsonString) => {
-                    Ok(Some(UpdateData::JsonRows(String::from_utf8(vec)?)))
-                },
-                Some(TableReadFormat::ColumnsString) => {
-                    Ok(Some(UpdateData::JsonColumns(String::from_utf8(vec)?)))
-                },
-                None | Some(TableReadFormat::Arrow) => Ok(Some(UpdateData::Arrow(vec.into()))),
-            }
-        } else if let Ok(pystring) = input.downcast_bound::<PyString>(py) {
-            let string = pystring.extract::<String>()?;
-            match format {
-                None | Some(TableReadFormat::Csv) => Ok(Some(UpdateData::Csv(string))),
-                Some(TableReadFormat::JsonString) => Ok(Some(UpdateData::JsonRows(string))),
-                Some(TableReadFormat::ColumnsString) => Ok(Some(UpdateData::JsonColumns(string))),
-                Some(TableReadFormat::Arrow) => {
-                    Ok(Some(UpdateData::Arrow(string.into_bytes().into())))
-                },
-            }
-        } else if let Ok(pylist) = input.downcast_bound::<PyList>(py) {
-            let json_module = PyModule::import_bound(py, "json")?;
-            let string = json_module.call_method("dumps", (pylist,), None)?;
-            Ok(Some(UpdateData::JsonRows(string.extract::<String>()?)))
-        } else if let Ok(pydict) = input.downcast_bound::<PyDict>(py) {
-            if pydict.keys().is_empty() {
-                return Err(PyValueError::new_err("Cannot infer type of empty dict"));
-            }
-
-            let first_key = pydict.keys().get_item(0)?;
-            let first_item = pydict
-                .get_item(first_key)?
-                .ok_or_else(|| PyValueError::new_err("Bad Input"))?;
-
-            if first_item.downcast::<PyList>().is_ok() {
-                let json_module = PyModule::import_bound(py, "json")?;
-                let string = json_module.call_method("dumps", (pydict,), None)?;
-                Ok(Some(UpdateData::JsonColumns(string.extract::<String>()?)))
-            } else {
-                Ok(None)
-            }
-        } else {
-            Ok(None)
-        }
-    }
-
-    fn from_py(
-        py: Python<'_>,
-        input: &Py<PyAny>,
-        format: Option<TableReadFormat>,
-    ) -> Result<UpdateData, PyErr> {
-        if let Some(x) = Self::from_py_partial(py, input, format)? {
-            Ok(x)
-        } else {
-            Err(PyValueError::new_err(format!(
-                "Unknown input type {:?}",
-                input.type_id()
-            )))
-        }
-    }
-}
-
-#[extend::ext]
-impl TableData {
-    fn from_py(
-        py: Python<'_>,
-        input: Py<PyAny>,
-        format: Option<TableReadFormat>,
-    ) -> Result<TableData, PyErr> {
-        if let Some(update) = UpdateData::from_py_partial(py, &input, format)? {
-            Ok(TableData::Update(update))
-        } else if let Ok(pylist) = input.downcast_bound::<PyList>(py) {
-            let json_module = PyModule::import_bound(py, "json")?;
-            let string = json_module.call_method("dumps", (pylist,), None)?;
-            Ok(UpdateData::JsonRows(string.extract::<String>()?).into())
-        } else if let Ok(pydict) = input.downcast_bound::<PyDict>(py) {
-            let first_key = pydict.keys().get_item(0)?;
-            let first_item = pydict
-                .get_item(first_key)?
-                .ok_or_else(|| PyValueError::new_err("Bad Input"))?;
-            if first_item.downcast::<PyList>().is_ok() {
-                let json_module = PyModule::import_bound(py, "json")?;
-                let string = json_module.call_method("dumps", (pydict,), None)?;
-                Ok(UpdateData::JsonColumns(string.extract::<String>()?).into())
-            } else {
-                let mut schema = vec![];
-                for (key, val) in pydict.into_iter() {
-                    schema.push((
-                        key.extract::<String>()?,
-                        val.extract::<String>()?.as_str().try_into().into_pyerr()?,
-                    ));
-                }
-
-                Ok(TableData::Schema(schema))
-            }
-        } else {
-            Err(PyValueError::new_err(format!(
-                "Unknown input type {:?}",
-                input.type_id()
-            )))
-        }
-    }
-}
-
 impl PyClient {
     pub fn new(handle_request: Py<PyAny>, handle_close: Option<Py<PyAny>>) -> Self {
         let client = Client::new_with_callback({
diff --git a/rust/perspective-python/src/client/table_data.rs b/rust/perspective-python/src/client/table_data.rs
@@ -0,0 +1,84 @@
+// ┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
+// ┃ ██████ ██████ ██████       █      █      █      █      █ █▄  ▀███ █       ┃
+// ┃ ▄▄▄▄▄█ █▄▄▄▄▄ ▄▄▄▄▄█  ▀▀▀▀▀█▀▀▀▀▀ █ ▀▀▀▀▀█ ████████▌▐███ ███▄  ▀█ █ ▀▀▀▀▀ ┃
+// ┃ █▀▀▀▀▀ █▀▀▀▀▀ █▀██▀▀ ▄▄▄▄▄ █ ▄▄▄▄▄█ ▄▄▄▄▄█ ████████▌▐███ █████▄   █ ▄▄▄▄▄ ┃
+// ┃ █      ██████ █  ▀█▄       █ ██████      █      ███▌▐███ ███████▄ █       ┃
+// ┣━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┫
+// ┃ Copyright (c) 2017, the Perspective Authors.                              ┃
+// ┃ ╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌ ┃
+// ┃ This file is part of the Perspective library, distributed under the terms ┃
+// ┃ of the [Apache License 2.0](https://www.apache.org/licenses/LICENSE-2.0). ┃
+// ┗━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┛
+
+use perspective_client::{ColumnType, TableData, TableReadFormat, UpdateData};
+use pyo3::exceptions::{PyTypeError, PyValueError};
+use pyo3::prelude::*;
+use pyo3::types::{PyAny, PyAnyMethods, PyDict, PyList, PyString, PyType};
+
+use super::update_data::UpdateDataExt;
+use crate::py_err::ResultTClientErrorExt;
+
+fn psp_type_from_py_type(_py: Python<'_>, val: Bound<'_, PyAny>) -> PyResult<ColumnType> {
+    if val.is_instance_of::<PyString>() {
+        val.extract::<String>()?.as_str().try_into().into_pyerr()
+    } else if let Ok(val) = val.downcast::<PyType>() {
+        match val.name()?.as_ref() {
+            "builtins.int" | "int" => Ok(ColumnType::Integer),
+            "builtins.float" | "float" => Ok(ColumnType::Float),
+            "builtins.str" | "str" => Ok(ColumnType::String),
+            "builtins.bool" | "bool" => Ok(ColumnType::Boolean),
+            "datetime.date" => Ok(ColumnType::Date),
+            "datetime.datetime" => Ok(ColumnType::Datetime),
+            type_name => Err(PyTypeError::new_err(type_name.to_string())),
+        }
+    } else {
+        Err(PyTypeError::new_err(format!(
+            "Unknown schema type {:?}",
+            val.get_type().name()?
+        )))
+    }
+}
+
+fn from_dict(py: Python<'_>, pydict: &Bound<'_, PyDict>) -> Result<TableData, PyErr> {
+    let first_key = pydict.keys().get_item(0)?;
+    let first_item = pydict
+        .get_item(first_key)?
+        .ok_or_else(|| PyValueError::new_err("Schema has no columns"))?;
+
+    if first_item.downcast::<PyList>().is_ok() {
+        let json_module = PyModule::import_bound(py, "json")?;
+        let string = json_module.call_method("dumps", (pydict,), None)?;
+        Ok(UpdateData::JsonColumns(string.extract::<String>()?).into())
+    } else {
+        let mut schema = vec![];
+        for (key, val) in pydict.into_iter() {
+            schema.push((key.extract::<String>()?, psp_type_from_py_type(py, val)?));
+        }
+
+        Ok(TableData::Schema(schema))
+    }
+}
+
+#[extend::ext]
+pub impl TableData {
+    fn from_py(
+        py: Python<'_>,
+        input: Py<PyAny>,
+        format: Option<TableReadFormat>,
+    ) -> Result<TableData, PyErr> {
+        if let Some(update) = UpdateData::from_py_partial(py, &input, format)? {
+            Ok(TableData::Update(update))
+        } else if let Ok(pylist) = input.downcast_bound::<PyList>(py) {
+            let json_module = PyModule::import_bound(py, "json")?;
+            let string = json_module.call_method("dumps", (pylist,), None)?;
+            Ok(UpdateData::JsonRows(string.extract::<String>()?).into())
+        } else if let Ok(pydict) = input.downcast_bound::<PyDict>(py) {
+            from_dict(py, pydict)
+        } else {
+            Err(PyTypeError::new_err(format!(
+                "Unknown input type {:?}",
+                input.bind(py).get_type().name()?
+            )))
+        }
+    }
+}
diff --git a/rust/perspective-python/src/client/update_data.rs b/rust/perspective-python/src/client/update_data.rs
diff --git a/rust/perspective-python/src/lib.rs b/rust/perspective-python/src/lib.rs
diff --git a/rust/perspective-python/src/py_err.rs b/rust/perspective-python/src/py_err.rs

Original file line number	Diff line number	Diff line change
`@@ -16,7 +16,6 @@`
`16`	`16`	`"Server",`
`17`	`17`	`"Client",`
`18`	`18`	`"PerspectiveError",`
`19`		`- "PerspectiveWidget",`
`20`	`19`	`"ProxySession",`
`21`	`20`	`]`
`22`	`21`