Type checking with pyright (#164)

* lower and upper more consistently

* one more

* handle bounds/bins/counts the same way

* lots of reactive dicts, but the UI has not changed

* data dump on the results page

* add a pragma: no cover

* reset widget values after checkbox change

* do not clean up values

* put tooltips in labels

* pull warning up to analysis panel. TODO: conditional

* move warning to bottom of list

* analysis definition JSON

* stubs for python

* stub a script on results page

* include column info in generated script

* closer to a runable notebook

* stuck on split_by_weight... maybe a library bug?

* margin stubs

* format python identifiers correctly

* script has gotten longer: does not make sense to check for exact equality

* fix syntactic problems in generated code

* fill in columns, but still WIP

* fix column names; tests pass

* move confidence

* simplify download panel

* add markdown cells

* tidy up

* switch requirements from mypy to pyright

* run pyright; currently fails

* fix pyright errors

* type input, output, session

* rm unused session

* more typing

* enable strict type checking; lots of errors!

* more typing. Add "finish()" for Templates
(Alternative is to implement __buffer__.)

* add a lot of ignores, but type checks pass locally

* I think Optional is needed under 3.9

* Remove strict, and most of the ignores

* tests in code generation

dp_wizard/app/__init__.py

@@ -1,9 +1,9 @@
 from pathlib import Path
 import logging
 
-from shiny import App, ui, reactive
+from shiny import App, ui, reactive, Inputs, Outputs, Session
 
-from dp_wizard.utils.argparse_helpers import get_cli_info
+from dp_wizard.utils.argparse_helpers import get_cli_info, CLIInfo
 from dp_wizard.app import analysis_panel, dataset_panel, results_panel, feedback_panel
 
 
@@ -26,16 +26,17 @@ def ctrl_c_reminder():  # pragma: no cover
     print("Session ended (Press CTRL+C to quit)")
 
 
-def make_server_from_cli_info(cli_info):
-    def server(input, output, session):  # pragma: no cover
-        csv_path = reactive.value(cli_info.csv_path)
+def make_server_from_cli_info(cli_info: CLIInfo):
+    def server(input: Inputs, output: Outputs, session: Session):  # pragma: no cover
+        cli_csv_path = cli_info.csv_path
+        csv_path = reactive.value("" if cli_csv_path is None else cli_csv_path)
         contributions = reactive.value(cli_info.contributions)
 
         lower_bounds = reactive.value({})
         upper_bounds = reactive.value({})
         bin_counts = reactive.value({})
         weights = reactive.value({})
-        epsilon = reactive.value(1)
+        epsilon = reactive.value(1.0)
 
         dataset_panel.dataset_server(
             input,

dp_wizard/app/analysis_panel.py

@@ -1,6 +1,7 @@
 from math import pow
+from typing import Iterable, Any
 
-from shiny import ui, reactive, render, req
+from shiny import ui, reactive, render, req, Inputs, Outputs, Session
 
 from dp_wizard.app.components.inputs import log_slider
 from dp_wizard.app.components.column_module import column_ui, column_server
@@ -39,7 +40,9 @@ def analysis_ui():
     )
 
 
-def _cleanup_reactive_dict(reactive_dict, keys_to_keep):  # pragma: no cover
+def _cleanup_reactive_dict(
+    reactive_dict: reactive.Value[dict[str, Any]], keys_to_keep: Iterable[str]
+):  # pragma: no cover
     reactive_dict_copy = {**reactive_dict()}
     keys_to_del = set(reactive_dict_copy.keys()) - set(keys_to_keep)
     for key in keys_to_del:
@@ -48,17 +51,17 @@ def _cleanup_reactive_dict(reactive_dict, keys_to_keep):  # pragma: no cover
 
 
 def analysis_server(
-    input,
-    output,
-    session,
-    csv_path,
-    contributions,
-    is_demo,
-    lower_bounds,
-    upper_bounds,
-    bin_counts,
-    weights,
-    epsilon,
+    input: Inputs,
+    output: Outputs,
+    session: Session,
+    csv_path: reactive.Value[str],
+    contributions: reactive.Value[int],
+    is_demo: bool,
+    lower_bounds: reactive.Value[dict[str, float]],
+    upper_bounds: reactive.Value[dict[str, float]],
+    bin_counts: reactive.Value[dict[str, int]],
+    weights: reactive.Value[dict[str, str]],
+    epsilon: reactive.Value[float],
 ):  # pragma: no cover
     @reactive.calc
     def button_enabled():
@@ -133,7 +136,7 @@ def columns_ui():
                             """
                             )
                         ],
-                        col_widths=col_widths,
+                        col_widths=col_widths,  # type: ignore
                     )
                     if column_ids
                     else []

dp_wizard/app/components/column_module.py

@@ -1,21 +1,21 @@
 from logging import info
 
-from shiny import ui, render, module, reactive
+from shiny import ui, render, module, reactive, Inputs, Outputs, Session
 
 from dp_wizard.utils.dp_helper import make_confidence_accuracy_histogram
 from dp_wizard.utils.shared import plot_histogram
 from dp_wizard.utils.code_generators import make_column_config_block
 from dp_wizard.app.components.outputs import output_code_sample, demo_tooltip
 
 
-default_weight = 2
+default_weight = "2"
 
 col_widths = {
     # Controls stay roughly a constant width;
     # Graph expands to fill space.
-    "sm": (4, 8),
-    "md": (3, 9),
-    "lg": (2, 10),
+    "sm": [4, 8],
+    "md": [3, 9],
+    "lg": [2, 10],
 }
 
 
@@ -37,9 +37,9 @@ def column_ui():  # pragma: no cover
                 "weight",
                 ["Weight", ui.output_ui("weight_tooltip_ui")],
                 choices={
-                    1: "Less accurate",
+                    "1": "Less accurate",
                     default_weight: "Default",
-                    4: "More accurate",
+                    "4": "More accurate",
                 },
                 selected=default_weight,
                 width=width,
@@ -50,31 +50,31 @@ def column_ui():  # pragma: no cover
             # Make plot smaller than default: about the same size as the other column.
             output_code_sample("Column Definition", "column_code"),
         ],
-        col_widths=col_widths,
+        col_widths=col_widths,  # type: ignore
     )
 
 
 @module.server
 def column_server(
-    input,
-    output,
-    session,
-    name,
-    contributions,
-    epsilon,
-    lower_bounds,
-    upper_bounds,
-    bin_counts,
-    weights,
-    is_demo,
+    input: Inputs,
+    output: Outputs,
+    session: Session,
+    name: str,
+    contributions: int,
+    epsilon: float,
+    lower_bounds: reactive.Value[dict[str, float]],
+    upper_bounds: reactive.Value[dict[str, float]],
+    bin_counts: reactive.Value[dict[str, int]],
+    weights: reactive.Value[dict[str, str]],
+    is_demo: bool,
 ):  # pragma: no cover
     @reactive.effect
     def _set_all_inputs():
         with reactive.isolate():  # Without isolate, there is an infinite loop.
             ui.update_numeric("lower", value=lower_bounds().get(name, 0))
             ui.update_numeric("upper", value=upper_bounds().get(name, 10))
             ui.update_numeric("bins", value=bin_counts().get(name, 10))
-            ui.update_numeric("weight", value=weights().get(name, default_weight))
+            ui.update_numeric("weight", value=int(weights().get(name, default_weight)))
 
     @reactive.effect
     @reactive.event(input.lower)
@@ -89,12 +89,12 @@ def _set_upper():
     @reactive.effect
     @reactive.event(input.bins)
     def _set_bins():
-        bin_counts.set({**bin_counts(), name: float(input.bins())})
+        bin_counts.set({**bin_counts(), name: int(input.bins())})
 
     @reactive.effect
     @reactive.event(input.weight)
     def _set_weight():
-        weights.set({**weights(), name: float(input.weight())})
+        weights.set({**weights(), name: input.weight()})
 
     @render.ui
     def bounds_tooltip_ui():
@@ -149,7 +149,7 @@ def column_plot():
         upper_x = float(input.upper())
         bin_count = int(input.bins())
         weight = float(input.weight())
-        weights_sum = sum(weights().values())
+        weights_sum = sum(float(weight) for weight in weights().values())
         info(f"Weight ratio for {name}: {weight}/{weights_sum}")
         if weights_sum == 0:
             # This function is triggered when column is removed;

dp_wizard/app/components/inputs.py

@@ -2,7 +2,7 @@
 from shiny import ui
 
 
-def log_slider(id, lower, upper):
+def log_slider(id: str, lower: float, upper: float):
     # Rather than engineer a new widget, hide the numbers we don't want.
     # The rendered widget doesn't have a unique ID, but the following
     # element does, so we can use some fancy CSS to get the preceding element.

dp_wizard/app/components/outputs.py

@@ -3,14 +3,14 @@
 from faicons import icon_svg
 
 
-def output_code_sample(title, name_of_render_function):
+def output_code_sample(title: str, name_of_render_function: str):
     return details(
         summary(f"Code sample: {title}"),
         ui.output_code(name_of_render_function),
     )
 
 
-def demo_tooltip(is_demo, text):  # pragma: no cover
+def demo_tooltip(is_demo: bool, text: str):  # pragma: no cover
     if is_demo:
         return ui.tooltip(
             icon_svg("circle-question"),

dp_wizard/app/dataset_panel.py

@@ -1,6 +1,6 @@
 from pathlib import Path
 
-from shiny import ui, reactive, render
+from shiny import ui, reactive, render, Inputs, Outputs, Session
 
 from dp_wizard.utils.argparse_helpers import get_cli_info
 from dp_wizard.app.components.outputs import output_code_sample, demo_tooltip
@@ -9,9 +9,7 @@
 
 def dataset_ui():
     cli_info = get_cli_info()
-    csv_placeholder = (
-        None if cli_info.csv_path is None else Path(cli_info.csv_path).name
-    )
+    csv_placeholder = "" if cli_info.csv_path is None else Path(cli_info.csv_path).name
 
     return ui.nav_panel(
         "Select Dataset",
@@ -41,7 +39,12 @@ def dataset_ui():
 
 
 def dataset_server(
-    input, output, session, csv_path=None, contributions=None, is_demo=None
+    input: Inputs,
+    output: Outputs,
+    session: Session,
+    csv_path: reactive.Value[str],
+    contributions: reactive.Value[int],
+    is_demo: bool,
 ):  # pragma: no cover
     @reactive.effect
     @reactive.event(input.csv_path)

dp_wizard/app/feedback_panel.py

@@ -1,4 +1,4 @@
-from shiny import ui
+from shiny import ui, Inputs, Outputs, Session
 from htmltools import HTML
 
 
@@ -25,8 +25,8 @@ def feedback_ui():
 
 
 def feedback_server(
-    input,
-    output,
-    session,
+    input: Inputs,
+    output: Outputs,
+    session: Session,
 ):  # pragma: no cover
     pass

dp_wizard/app/results_panel.py

@@ -1,4 +1,4 @@
-from shiny import ui, render, reactive
+from shiny import ui, render, reactive, Inputs, Outputs, Session
 
 from dp_wizard.utils.code_generators import (
     NotebookGenerator,
@@ -26,16 +26,16 @@ def results_ui():
 
 
 def results_server(
-    input,
-    output,
-    session,
-    csv_path,
-    contributions,
-    lower_bounds,
-    upper_bounds,
-    bin_counts,
-    weights,
-    epsilon,
+    input: Inputs,
+    output: Outputs,
+    session: Session,
+    csv_path: reactive.Value[str],
+    contributions: reactive.Value[int],
+    lower_bounds: reactive.Value[dict[str, float]],
+    upper_bounds: reactive.Value[dict[str, float]],
+    bin_counts: reactive.Value[dict[str, int]],
+    weights: reactive.Value[dict[str, str]],
+    epsilon: reactive.Value[float],
 ):  # pragma: no cover
     @reactive.calc
     def analysis_plan() -> AnalysisPlan:

dp_wizard/utils/argparse_helpers.py

@@ -4,10 +4,10 @@
 import csv
 import random
 from warnings import warn
-from collections import namedtuple
+from typing import NamedTuple, Optional
 
 
-def _existing_csv_type(arg):
+def _existing_csv_type(arg: str) -> Path:
     path = Path(arg)
     if not path.exists():
         raise ArgumentTypeError(f"No such file: {arg}")
@@ -54,7 +54,7 @@ def _get_args():
         return arg_parser.parse_args()  # pragma: no cover
 
 
-def _clip(n, lower, upper):
+def _clip(n: float, lower: float, upper: float) -> float:
     """
     >>> _clip(-5, 0, 10)
     0
@@ -66,7 +66,13 @@ def _clip(n, lower, upper):
     return max(min(n, upper), lower)
 
 
-def _get_demo_csv_contrib():
+class CLIInfo(NamedTuple):
+    csv_path: Optional[str]
+    contributions: int
+    is_demo: bool
+
+
+def _get_demo_csv_contrib() -> CLIInfo:
     """
     >>> csv_path, contributions, is_demo = _get_demo_csv_contrib()
     >>> with open(csv_path, newline="") as csv_handle:
@@ -103,10 +109,7 @@ def _get_demo_csv_contrib():
                     }
                 )
 
-    return CLIInfo(csv_path=csv_path, contributions=contributions, is_demo=True)
-
-
-CLIInfo = namedtuple("CLIInfo", ["csv_path", "contributions", "is_demo"])
+    return CLIInfo(csv_path=str(csv_path), contributions=contributions, is_demo=True)
 
 
 def get_cli_info():  # pragma: no cover