edsl 0.1.37.dev6__py3-none-any.whl → 0.1.38__py3-none-any.whl

edsl/results/ResultsGGMixin.py

@@ -1,121 +1,121 @@
-"""Mixin class for ggplot2 plotting."""
-
-import subprocess
-import tempfile
-from typing import Optional
-
-
-class ResultsGGMixin:
-    """Mixin class for ggplot2 plotting."""
-
-    def ggplot2(
-        self,
-        ggplot_code: str,
-        filename: str = None,
-        shape="wide",
-        sql: str = None,
-        remove_prefix: bool = True,
-        debug: bool = False,
-        height=4,
-        width=6,
-        format="svg",
-        factor_orders: Optional[dict] = None,
-    ):
-        """Create a ggplot2 plot from a DataFrame.
-
-        :param ggplot_code: The ggplot2 code to execute.
-        :param filename: The filename to save the plot to.
-        :param shape: The shape of the data in the DataFrame (wide or long).
-        :param sql: The SQL query to execute beforehand to manipulate the data.
-        :param remove_prefix: Whether to remove the prefix from the column names.
-        :param debug: Whether to print the R code instead of executing it.
-        :param height: The height of the plot in inches.
-        :param width: The width of the plot in inches.
-        :param format: The format to save the plot in (png or svg).
-        :param factor_orders: A dictionary of factor columns and their order.
-        """
-
-        if sql == None:
-            sql = "select * from self"
-
-        if shape == "long":
-            df = self.sql(sql, shape="long")
-        elif shape == "wide":
-            df = self.sql(sql, shape="wide", remove_prefix=remove_prefix)
-
-        # Convert DataFrame to CSV format
-        csv_data = df.to_csv(index=False)
-
-        # Embed the CSV data within the R script
-        csv_data_escaped = csv_data.replace("\n", "\\n").replace("'", "\\'")
-        read_csv_code = f"self <- read.csv(text = '{csv_data_escaped}', sep = ',')\n"
-
-        if factor_orders is not None:
-            for factor, order in factor_orders.items():
-                # read_csv_code += f"""self${{{factor}}} <- factor(self${{{factor}}}, levels=c({','.join(['"{}"'.format(x) for x in order])}))"""
-
-                level_string = ", ".join([f'"{x}"' for x in order])
-                read_csv_code += (
-                    f"self${factor} <- factor(self${factor}, levels=c({level_string}))"
-                )
-                read_csv_code += "\n"
-
-        # Load ggplot2 library
-        load_ggplot2 = "library(ggplot2)\n"
-
-        # Check if a filename is provided for the plot, if not create a temporary one
-        if not filename:
-            filename = tempfile.mktemp(suffix=f".{format}")
-
-        # Combine all R script parts
-        full_r_code = load_ggplot2 + read_csv_code + ggplot_code
-
-        # Add command to save the plot to a file
-        full_r_code += f'\nggsave("{filename}", plot = last_plot(), width = {width}, height = {height}, device = "{format}")'
-
-        if debug:
-            print(full_r_code)
-            return
-
-        result = subprocess.run(
-            ["Rscript", "-"],
-            input=full_r_code,
-            text=True,
-            stdout=subprocess.PIPE,
-            stderr=subprocess.PIPE,
-        )
-
-        if result.returncode != 0:
-            if result.returncode == 127:  # 'command not found'
-                raise RuntimeError(
-                    "Rscript is probably not installed. Please install R from https://cran.r-project.org/"
-                )
-            else:
-                raise RuntimeError(
-                    f"An error occurred while running Rscript: {result.stderr}"
-                )
-
-        if result.stderr:
-            print("Error in R script:", result.stderr)
-        else:
-            self._display_plot(filename, width, height)
-
-    def _display_plot(self, filename: str, width: float, height: float):
-        """Display the plot in the notebook."""
-        import matplotlib.pyplot as plt
-        import matplotlib.image as mpimg
-
-        if filename.endswith(".png"):
-            img = mpimg.imread(filename)
-            plt.figure(
-                figsize=(width, height)
-            )  # Set the figure size (width, height) in inches
-            plt.imshow(img)
-            plt.axis("off")
-            plt.show()
-        elif filename.endswith(".svg"):
-            from IPython.display import SVG, display
-
-            display(SVG(filename=filename))
-        else:
-            print("Unsupported file format. Please provide a PNG or SVG file.")
+"""Mixin class for ggplot2 plotting."""
+
+import subprocess
+import tempfile
+from typing import Optional
+
+
+class ResultsGGMixin:
+    """Mixin class for ggplot2 plotting."""
+
+    def ggplot2(
+        self,
+        ggplot_code: str,
+        filename: str = None,
+        shape="wide",
+        sql: str = None,
+        remove_prefix: bool = True,
+        debug: bool = False,
+        height=4,
+        width=6,
+        format="svg",
+        factor_orders: Optional[dict] = None,
+    ):
+        """Create a ggplot2 plot from a DataFrame.
+
+        :param ggplot_code: The ggplot2 code to execute.
+        :param filename: The filename to save the plot to.
+        :param shape: The shape of the data in the DataFrame (wide or long).
+        :param sql: The SQL query to execute beforehand to manipulate the data.
+        :param remove_prefix: Whether to remove the prefix from the column names.
+        :param debug: Whether to print the R code instead of executing it.
+        :param height: The height of the plot in inches.
+        :param width: The width of the plot in inches.
+        :param format: The format to save the plot in (png or svg).
+        :param factor_orders: A dictionary of factor columns and their order.
+        """
+
+        if sql == None:
+            sql = "select * from self"
+
+        if shape == "long":
+            df = self.sql(sql, shape="long")
+        elif shape == "wide":
+            df = self.sql(sql, shape="wide", remove_prefix=remove_prefix)
+
+        # Convert DataFrame to CSV format
+        csv_data = df.to_csv(index=False)
+
+        # Embed the CSV data within the R script
+        csv_data_escaped = csv_data.replace("\n", "\\n").replace("'", "\\'")
+        read_csv_code = f"self <- read.csv(text = '{csv_data_escaped}', sep = ',')\n"
+
+        if factor_orders is not None:
+            for factor, order in factor_orders.items():
+                # read_csv_code += f"""self${{{factor}}} <- factor(self${{{factor}}}, levels=c({','.join(['"{}"'.format(x) for x in order])}))"""
+
+                level_string = ", ".join([f'"{x}"' for x in order])
+                read_csv_code += (
+                    f"self${factor} <- factor(self${factor}, levels=c({level_string}))"
+                )
+                read_csv_code += "\n"
+
+        # Load ggplot2 library
+        load_ggplot2 = "library(ggplot2)\n"
+
+        # Check if a filename is provided for the plot, if not create a temporary one
+        if not filename:
+            filename = tempfile.mktemp(suffix=f".{format}")
+
+        # Combine all R script parts
+        full_r_code = load_ggplot2 + read_csv_code + ggplot_code
+
+        # Add command to save the plot to a file
+        full_r_code += f'\nggsave("{filename}", plot = last_plot(), width = {width}, height = {height}, device = "{format}")'
+
+        if debug:
+            print(full_r_code)
+            return
+
+        result = subprocess.run(
+            ["Rscript", "-"],
+            input=full_r_code,
+            text=True,
+            stdout=subprocess.PIPE,
+            stderr=subprocess.PIPE,
+        )
+
+        if result.returncode != 0:
+            if result.returncode == 127:  # 'command not found'
+                raise RuntimeError(
+                    "Rscript is probably not installed. Please install R from https://cran.r-project.org/"
+                )
+            else:
+                raise RuntimeError(
+                    f"An error occurred while running Rscript: {result.stderr}"
+                )
+
+        if result.stderr:
+            print("Error in R script:", result.stderr)
+        else:
+            self._display_plot(filename, width, height)
+
+    def _display_plot(self, filename: str, width: float, height: float):
+        """Display the plot in the notebook."""
+        import matplotlib.pyplot as plt
+        import matplotlib.image as mpimg
+
+        if filename.endswith(".png"):
+            img = mpimg.imread(filename)
+            plt.figure(
+                figsize=(width, height)
+            )  # Set the figure size (width, height) in inches
+            plt.imshow(img)
+            plt.axis("off")
+            plt.show()
+        elif filename.endswith(".svg"):
+            from IPython.display import SVG, display
+
+            display(SVG(filename=filename))
+        else:
+            print("Unsupported file format. Please provide a PNG or SVG file.")

edsl/results/ResultsToolsMixin.py

@@ -1,98 +1,98 @@
-class ResultsToolsMixin:
-    def get_themes(
-        self,
-        field: str,
-        context: str,
-        max_values=100,
-        num_themes: int = 10,
-        seed=None,
-        progress_bar=False,
-        print_exceptions=False,
-    ) -> list:
-        values = [
-            str(txt)[:1000]
-            for txt in self.shuffle(seed=seed).select(field).to_list()[:max_values]
-        ]
-        from edsl import ScenarioList
-        from edsl.questions import QuestionList, QuestionCheckBox
-
-        q = QuestionList(
-            question_text=f"""
-        {context}
-        Here are some examples: { values }. 
-        What are some 5-8 word themes that would mostly capture these examples?
-        Please shoot for {num_themes} as target number of themes.
-        """,
-            question_name="themes",
-        )
-        results = q.run(print_exceptions=print_exceptions, progress_bar=progress_bar)
-        return results.select("themes").first()
-
-    def answers_to_themes(
-        self,
-        field,
-        context: str,
-        themes: list,
-        progress_bar=False,
-        print_exceptions=False,
-    ) -> dict:
-        from edsl import ScenarioList
-        from edsl import QuestionCheckBox
-
-        values = self.select(field).to_list()
-        scenarios = ScenarioList.from_list("field", values).add_value(
-            "context", context
-        )
-        q = QuestionCheckBox(
-            question_text="""
-            {{ context }}
-            Consider the following response: 
-
-            " {{ field }} "
-
-            Please check the themes that apply to these examples.
-            If blank, please check 'None'.
-            """,
-            question_options=themes + ["None", "Other"],
-            question_name="themes",
-        )
-        results = q.by(scenarios).run(
-            progress_bar=progress_bar, print_exceptions=print_exceptions
-        )
-        return {k: v for k, v in results.select("field", "themes").to_list()}
-
-    def apply_themes(self, field: str, new_field: str, answers_to_themes: dict):
-        def translate(x):
-            return answers_to_themes.get(x, "Other")
-
-        self.mutate(f"{new_field} = f({field})", functions_dict={"f": translate})
-        return self
-
-    def auto_theme(
-        self,
-        field: str,
-        context: str,
-        themes: list[str],
-        newfield: str = None,
-        progress_bar=False,
-        print_exceptions=False,
-    ) -> tuple:
-        """
-        :param field: The field to be themed.
-        :param context: The context of the field.
-        :param themes: The list of themes.
-        :param newfield: The new field name.
-
-        """
-
-        if not newfield:
-            newfield = f"{field}_themes"
-
-        answers_to_themes = self.answers_to_themes(
-            field=field,
-            context=context,
-            themes=themes,
-            progress_bar=progress_bar,
-            print_exceptions=print_exceptions,
-        )
-        return self.apply_themes(field, newfield, answers_to_themes), themes
+class ResultsToolsMixin:
+    def get_themes(
+        self,
+        field: str,
+        context: str,
+        max_values=100,
+        num_themes: int = 10,
+        seed=None,
+        progress_bar=False,
+        print_exceptions=False,
+    ) -> list:
+        values = [
+            str(txt)[:1000]
+            for txt in self.shuffle(seed=seed).select(field).to_list()[:max_values]
+        ]
+        from edsl import ScenarioList
+        from edsl.questions import QuestionList, QuestionCheckBox
+
+        q = QuestionList(
+            question_text=f"""
+        {context}
+        Here are some examples: { values }. 
+        What are some 5-8 word themes that would mostly capture these examples?
+        Please shoot for {num_themes} as target number of themes.
+        """,
+            question_name="themes",
+        )
+        results = q.run(print_exceptions=print_exceptions, progress_bar=progress_bar)
+        return results.select("themes").first()
+
+    def answers_to_themes(
+        self,
+        field,
+        context: str,
+        themes: list,
+        progress_bar=False,
+        print_exceptions=False,
+    ) -> dict:
+        from edsl import ScenarioList
+        from edsl import QuestionCheckBox
+
+        values = self.select(field).to_list()
+        scenarios = ScenarioList.from_list("field", values).add_value(
+            "context", context
+        )
+        q = QuestionCheckBox(
+            question_text="""
+            {{ context }}
+            Consider the following response: 
+
+            " {{ field }} "
+
+            Please check the themes that apply to these examples.
+            If blank, please check 'None'.
+            """,
+            question_options=themes + ["None", "Other"],
+            question_name="themes",
+        )
+        results = q.by(scenarios).run(
+            progress_bar=progress_bar, print_exceptions=print_exceptions
+        )
+        return {k: v for k, v in results.select("field", "themes").to_list()}
+
+    def apply_themes(self, field: str, new_field: str, answers_to_themes: dict):
+        def translate(x):
+            return answers_to_themes.get(x, "Other")
+
+        self.mutate(f"{new_field} = f({field})", functions_dict={"f": translate})
+        return self
+
+    def auto_theme(
+        self,
+        field: str,
+        context: str,
+        themes: list[str],
+        newfield: str = None,
+        progress_bar=False,
+        print_exceptions=False,
+    ) -> tuple:
+        """
+        :param field: The field to be themed.
+        :param context: The context of the field.
+        :param themes: The list of themes.
+        :param newfield: The new field name.
+
+        """
+
+        if not newfield:
+            newfield = f"{field}_themes"
+
+        answers_to_themes = self.answers_to_themes(
+            field=field,
+            context=context,
+            themes=themes,
+            progress_bar=progress_bar,
+            print_exceptions=print_exceptions,
+        )
+        return self.apply_themes(field, newfield, answers_to_themes), themes