clear-skies 2.0.5__py3-none-any.whl → 2.0.7__py3-none-any.whl

clearskies/endpoints/advanced_search.py

@@ -0,0 +1,526 @@
+from __future__ import annotations
+
+import inspect
+from collections import OrderedDict
+from typing import TYPE_CHECKING, Any, Type
+
+import clearskies.configs
+import clearskies.exceptions
+from clearskies import authentication, autodoc, typing
+from clearskies.endpoints.simple_search import SimpleSearch
+from clearskies.functional import string
+from clearskies.input_outputs import InputOutput
+
+if TYPE_CHECKING:
+    from clearskies import SecurityHeader
+    from clearskies.model import Model
+
+
+class AdvancedSearch(SimpleSearch):
+    """
+    An endpoint that grants the client extensive control over searching and filtering.
+
+    Rather than accepting URL parameters (like the SimpleSearch endpoint), this endpoint accepts a JSON POST
+    body.  Search conditions are specified as a list of dictionaries containing `column`, `operator`, and
+    `value`.  It also accepts up to two sort directives.  Of course, while this endpoint supports arbitrary
+    searching, it won't work if the backend itself doesn't support it.  The following is the list of allowed
+    keys in the JSON body:
+
+    | Name  | Type                 | Description                                                                | Example |
+    |-------|----------------------|----------------------------------------------------------------------------|---------|
+    | sort  | list[dict[str, str]] | A list of sort directives containing `column` and `direction`              | `{"sort": [ {"column": "age", "direction": "desc} ] }` |
+    | limit | int                  | The number of records to return                                            | `{"limit": `100`}` |
+    | where | list[dict[str, Any]] | A list of conditions containing `column`, `operator`, and `value`          | `{"where": [ {"column": "age", "operator": ">", "value": 10} ] }` |
+    | *     | str, int             | Pagination information.  The key name and value type depend on the backend | `{"start": 100}` |
+
+    Here's an example making use of the AdvancedSearch endpoint:
+
+    ```python
+    import clearskies
+
+
+    class Company(clearskies.Model):
+        id_column_name = "id"
+        backend = clearskies.backends.MemoryBackend()
+
+        id = clearskies.columns.Uuid()
+        name = clearskies.columns.String()
+
+
+    class User(clearskies.Model):
+        id_column_name = "id"
+        backend = clearskies.backends.MemoryBackend()
+
+        id = clearskies.columns.Uuid()
+        name = clearskies.columns.String()
+        username = clearskies.columns.String()
+        age = clearskies.columns.Integer()
+        company_id = clearskies.columns.BelongsToId(Company, readable_parent_columns=["id", "name"])
+        company = clearskies.columns.BelongsToModel("company_id")
+
+
+    wsgi = clearskies.contexts.WsgiRef(
+        clearskies.endpoints.AdvancedSearch(
+            model_class=User,
+            readable_column_names=["id", "name", "username", "age", "company"],
+            sortable_column_names=["name", "username", "age", "company.name"],
+            searchable_column_names=["id", "name", "username", "age", "company_id", "company.name"],
+            default_sort_column_name="name",
+        ),
+        bindings={
+            "memory_backend_default_data": [
+                {
+                    "model_class": Company,
+                    "records": [
+                        {"id": "5-5-5-5", "name": "Bob's Widgets"},
+                        {"id": "3-3-3-3", "name": "New Venture"},
+                        {"id": "7-7-7-7", "name": "Jane's Cool Stuff"},
+                    ],
+                },
+                {
+                    "model_class": User,
+                    "records": [
+                        {
+                            "id": "1-2-3-4",
+                            "name": "Bob Brown",
+                            "username": "bobbrown",
+                            "age": 18,
+                            "company_id": "5-5-5-5",
+                        },
+                        {
+                            "id": "1-2-3-5",
+                            "name": "Jane Doe",
+                            "username": "janedoe",
+                            "age": 52,
+                            "company_id": "7-7-7-7",
+                        },
+                        {
+                            "id": "1-2-3-6",
+                            "name": "Greg",
+                            "username": "greg",
+                            "age": 37,
+                            "company_id": "7-7-7-7",
+                        },
+                        {
+                            "id": "1-2-3-7",
+                            "name": "Curious George",
+                            "username": "curious",
+                            "age": 7,
+                            "company_id": "3-3-3-3",
+                        },
+                    ],
+                },
+            ],
+        },
+    )
+    wsgi()
+    ```
+
+    If you invoke the endpoint without any additional data, it will simply list all records:
+
+    ```bash
+    $ curl 'http://localhost:8080/' | jq
+    {
+        "status": "success",
+        "error": "",
+        "data": [
+            {
+                "id": "1-2-3-4",
+                "name": "Bob Brown",
+                "username": "bobbrown",
+                "age": 18,
+                "company": {
+                    "id": "5-5-5-5",
+                    "name": "Bob's Widgets"
+                }
+            },
+            {
+                "id": "1-2-3-7",
+                "name": "Curious George",
+                "username": "curious",
+                "age": 7,
+                "company": {
+                    "id": "3-3-3-3",
+                    "name": "New Venture"
+                }
+            },
+            {
+                "id": "1-2-3-6",
+                "name": "Greg",
+                "username": "greg",
+                "age": 37,
+                "company": {
+                    "id": "7-7-7-7",
+                    "name": "Jane's Cool Stuff"
+                }
+            },
+            {
+                "id": "1-2-3-5",
+                "name": "Jane Doe",
+                "username": "janedoe",
+                "age": 52,
+                "company": {
+                    "id": "7-7-7-7",
+                    "name": "Jane's Cool Stuff"
+                }
+            }
+        ],
+        "pagination": {
+            "number_results": 4,
+            "limit": 50,
+            "next_page": {}
+        },
+        "input_errors": {}
+    }
+    ```
+
+    Of course you can also sort and paginate.  Keep in mind that pagination is backend-dependent:
+
+    ```bash
+    $ curl 'http://localhost:8080/' -d '{"sort":[ {"column": "name", "direction": "desc"} ], "limit": 2, "start": 1}' | jq
+    {
+        "status": "success",
+        "error": "",
+        "data": [
+            {
+                "id": "1-2-3-6",
+                "name": "Greg",
+                "username": "greg",
+                "age": 37,
+                "company": {
+                    "id": "7-7-7-7",
+                    "name": "Jane's Cool Stuff"
+                }
+            },
+            {
+                "id": "1-2-3-7",
+                "name": "Curious George",
+                "username": "curious",
+                "age": 7,
+                "company": {
+                    "id": "3-3-3-3",
+                    "name": "New Venture"
+                }
+            }
+        ],
+        "pagination": {
+            "number_results": 4,
+            "limit": 2,
+            "next_page": {
+            "start": 3
+            }
+        },
+        "input_errors": {}
+    }
+
+    ```
+
+    Note that sorting on columns in related models is done via the syntax `relationship_column.column_name`.  These
+    must be listed as such in the list of sortable/searchable columns, and then you use the same name to sort/search
+    by them:
+
+    ```bash
+    $ curl 'http://localhost:8080/' -d '{"sort":[ {"column": "company.name", "direction": "desc"}, {"column": "age", "direction": "asc"} ]}' | jq
+    {
+        "status": "success",
+        "error": "",
+        "data": [
+            {
+                "id": "1-2-3-7",
+                "name": "Curious George",
+                "username": "curious",
+                "age": 7,
+                "company": {
+                    "id": "3-3-3-3",
+                    "name": "New Venture"
+                }
+            },
+                {
+                "id": "1-2-3-6",
+                "name": "Greg",
+                "username": "greg",
+                "age": 37,
+                "company": {
+                    "id": "7-7-7-7",
+                    "name": "Jane's Cool Stuff"
+                }
+            },
+                {
+                "id": "1-2-3-5",
+                "name": "Jane Doe",
+                "username": "janedoe",
+                "age": 52,
+                "company": {
+                    "id": "7-7-7-7",
+                    "name": "Jane's Cool Stuff"
+                }
+            },
+                {
+                "id": "1-2-3-4",
+                "name": "Bob Brown",
+                "username": "bobbrown",
+                "age": 18,
+                "company": {
+                    "id": "5-5-5-5",
+                    "name": "Bob's Widgets"
+                }
+            }
+        ],
+        "pagination": {
+            "number_results": 4,
+            "limit": 50,
+            "next_page": {}
+        },
+        "input_errors": {}
+    }
+
+    ```
+
+    And finally searching:
+
+    ```bash
+    $ curl 'http://localhost:8080/' -d '{"where":[ {"column": "age", "operator": "<=", "value": 37}, {"column": "username", "operator": "in", "value": ["curious", "greg"]} ]}' | jq
+    {
+        "status": "success",
+        "error": "",
+        "data": [
+            {
+                "id": "1-2-3-7",
+                "name": "Curious George",
+                "username": "curious",
+                "age": 7,
+                "company": {
+                    "id": "3-3-3-3",
+                    "name": "New Venture"
+                }
+            },
+            {
+                "id": "1-2-3-6",
+                "name": "Greg",
+                "username": "greg",
+                "age": 37,
+                "company": {
+                    "id": "7-7-7-7",
+                    "name": "Jane's Cool Stuff"
+                }
+            }
+        ],
+        "pagination": {
+            "number_results": 2,
+            "limit": 50,
+            "next_page": {}
+        },
+        "input_errors": {}
+    }
+
+    ```
+
+    In terms of the allowed search operators, the standard list of operators is:
+
+     * `<=>`
+     * `!=`
+     * `<=`
+     * `>=`
+     * `>`
+     * `<`
+     * `=`
+     * `in`
+     * `is not null`
+     * `is null`
+     * `is not`
+     * `is`
+     * `like`
+
+    Although not all operators are supported by all columns.  You can use `%` with the `LIKE` operator
+    to perform a wildcard search.
+
+    """
+
+    @property
+    def allowed_request_keys(self) -> list[str]:
+        return self.internal_request_keys
+
+    @property
+    def internal_request_keys(self) -> list[str]:
+        return ["sort", "limit", "where"]
+
+    def check_request_data(
+        self, request_data: dict[str, Any], query_parameters: dict[str, Any], pagination_data: dict[str, Any]
+    ) -> None:
+        if pagination_data:
+            error = self.model.validate_pagination_data(pagination_data, self.auto_case_internal_column_name)
+            if error:
+                raise clearskies.exceptions.ClientError(error)
+        if query_parameters:
+            raise clearskies.exceptions.ClientError("Query parameters were found but are not supported.")
+        for key in request_data.keys():
+            if key not in self.allowed_request_keys:
+                raise clearskies.exceptions.ClientError(
+                    f"Invalid request parameter found in request body: '{key}'.  Expected parameters: "
+                    + ", ".join([self.auto_case_internal_column_name(key) for key in self.allowed_request_keys])
+                )
+        self.validate_limit(request_data, {})
+        sort_key_name = self.auto_case_internal_column_name("sort")
+        sort = request_data.get(sort_key_name, [])
+        if not isinstance(sort, list):
+            raise clearskies.exceptions.ClientError(
+                f"'{sort_key_name}' property in request body should be a list, but I found a value of type "
+                + sort.__class__.__name
+            )
+        if sort:
+            column_key_name = self.auto_case_internal_column_name("column")
+            direction_key_name = self.auto_case_internal_column_name("direction")
+            for index, sort_entry in enumerate(sort):
+                if not isinstance(sort_entry, dict):
+                    raise clearskies.exceptions.ClientError(
+                        f"'{sort_key_name}' should be a list of dictionaries, but entry #{index + 1} is a value of type '{sort_entry.__class__.__name}', not a dict"
+                    )
+                for key_name in [column_key_name, direction_key_name]:
+                    if not sort_entry.get(key_name):
+                        raise clearskies.exceptions.ClientError(
+                            f"Each entry in the sort list should contain both '{column_key_name}' and '{direction_key_name}' but entry #{index + 1} is missing '{key_name}'"
+                        )
+                    if not isinstance(sort_entry[key_name], str):
+                        raise clearskies.exceptions.ClientError(
+                            f"{key_name}' must be a string, but for entry #{index + 1} it is a value of type "
+                            + sort_entry[key_name].__class__.__name__
+                        )
+                if sort_entry[direction_key_name].lower() not in ["asc", "desc"]:
+                    raise clearskies.exceptions.ClientError(
+                        f"{direction_key_name}' must be either 'ASC' or 'DESC', but a different value was found for entry #{index + 1}"
+                    )
+                if self.auto_case_column_name(sort_entry[column_key_name], False) not in self.sortable_column_names:
+                    raise clearskies.exceptions.ClientError(
+                        f"Invalid sort column for entry #{index + 1}.  Allowed values are: "
+                        + ", ".join(
+                            [
+                                self.auto_case_column_name(column_name, False)
+                                for column_name in self.sortable_column_names
+                            ]
+                        )
+                    )
+        where_key_name = self.auto_case_internal_column_name("where")
+        where = request_data.get(where_key_name, [])
+        if not isinstance(where, list):
+            raise clearskies.exceptions.ClientError(
+                f"'{where_key_name}' property in request body should be a list, but I found a value of type "
+                + where.__class__.__name
+            )
+        if where:
+            column_key_name = self.auto_case_internal_column_name("column")
+            operator_key_name = self.auto_case_internal_column_name("operator")
+            value_key_name = self.auto_case_internal_column_name("value")
+            for index, where_entry in enumerate(where):
+                if not isinstance(where_entry, dict):
+                    raise clearskies.exceptions.ClientError(
+                        f"'{where_key_name}' should be a list of dictionaries, but entry #{index + 1} is a value of type '{where_entry.__class__.__name}', not a dict"
+                    )
+                for key_name in [column_key_name, operator_key_name, value_key_name]:
+                    if key_name not in where_entry:
+                        raise clearskies.exceptions.ClientError(
+                            f"Each entry in the where list should contain '{column_key_name}', '{operator_key_name}', and '{value_key_name}', but entry #{index + 1} is missing '{key_name}'"
+                        )
+                    if key_name != value_key_name and not isinstance(where_entry[key_name], str):
+                        raise clearskies.exceptions.ClientError(
+                            f"{key_name}' must be a string, but for entry #{index + 1} it is a value of type "
+                            + sort_entry[key_name].__class__.__name__
+                        )
+                    if where_entry[column_key_name] not in self.searchable_column_names:
+                        raise clearskies.exceptions.ClientError(
+                            f"Invalid where column for entry #{index + 1}.  Allowed values are: "
+                            + ", ".join(
+                                [
+                                    self.auto_case_column_name(column_name, True)
+                                    for column_name in self.searchable_column_names
+                                ]
+                            )
+                        )
+                    [relationship_column_name, column_name] = self.unpack_column_name_with_relationship(
+                        self.auto_case_column_name(where_entry[column_key_name], False),
+                    )
+                    operator = where_entry[operator_key_name].lower()
+                    value = where_entry[value_key_name]
+                    error_allowed_operators = None
+                    if relationship_column_name:
+                        column = self.columns[relationship_column_name]
+                        if not column.is_allowed_search_operator(operator, relationship_reference=column_name):
+                            error_allowed_operators = column.allowed_search_operators(
+                                relationship_reference=column_name
+                            )
+                        else:
+                            error = column.check_search_value(
+                                value if operator != "in" else value[0],
+                                where_entry[operator_key_name],
+                                relationship_reference=column_name,
+                            )
+                    else:
+                        column = self.columns[column_name]
+                        if not column.is_allowed_search_operator(operator):
+                            error_allowed_operators = column.allowed_search_operators()
+                        else:
+                            error = column.check_search_value(
+                                value if operator != "in" else value[0], where_entry[operator_key_name]
+                            )
+                    if error_allowed_operators:
+                        raise clearskies.exceptions.ClientError(
+                            f"Invalid operator for entry #{index + 1}.  Allowed operators are: "
+                            + ", ".join(column.allowed_search_operators(relationship_reference=column_name))
+                        )
+                    if error:
+                        raise clearskies.exceptions.ClientError(f"Invalid search value for entry #{index + 1}: {error}")
+
+    def configure_model_from_request_data(
+        self,
+        model: Model,
+        request_data: dict[str, Any],
+        query_parameters: dict[str, Any],
+        pagination_data: dict[str, Any],
+    ) -> Model:
+        if pagination_data:
+            model = model.pagination(**pagination_data)
+        sort = request_data.get(self.auto_case_internal_column_name("sort"), [])
+        if sort:
+            column_key_name = self.auto_case_internal_column_name("column")
+            direction_key_name = self.auto_case_internal_column_name("direction")
+            model = self.add_join(sort[0][column_key_name], model)
+            [primary_table_name, primary_column_name] = self.resolve_references_for_query(sort[0][column_key_name])
+            primary_direction = sort[0][direction_key_name]
+
+            if len(sort) > 1:
+                [secondary_table_name, secondary_column_name] = self.resolve_references_for_query(
+                    sort[1][column_key_name]
+                )
+                secondary_direction = sort[1][direction_key_name]
+            else:
+                secondary_column_name = ""
+                secondary_direction = ""
+                secondary_table_name = ""
+            model = model.sort_by(
+                primary_column_name if primary_column_name else "",
+                primary_direction if primary_direction else "",
+                primary_table_name=primary_table_name if primary_table_name else "",
+                secondary_column_name=secondary_column_name if secondary_column_name else "",
+                secondary_direction=secondary_direction if secondary_direction else "",
+                secondary_table_name=secondary_table_name if secondary_table_name else "",
+            )
+        if request_data.get("limit"):
+            model = model.limit(request_data["limit"])
+
+        for where in request_data.get(self.auto_case_internal_column_name("where"), []):
+            raw_column_name = self.auto_case_column_name(where[self.auto_case_internal_column_name("column")], False)
+            [relationship_column_name, column_name] = self.unpack_column_name_with_relationship(raw_column_name)
+            operator = where[self.auto_case_internal_column_name("operator")].lower()
+            value = where[self.auto_case_internal_column_name("value")]
+
+            model = self.add_join(raw_column_name, model)
+            if relationship_column_name:
+                model = self.columns[relationship_column_name].add_search(
+                    model, value, operator=operator, relationship_reference=column_name
+                )
+            else:
+                model = self.columns[column_name].add_search(model, value, operator=operator)
+
+        return model
+
+    def documentation_url_search_parameters(self) -> list[autodoc.request.Parameter]:
+        return []