Ensure clean flake8, pytest, and linter outputs.

Author: njhenry

pyproject.toml

@@ -49,6 +49,7 @@ disable = [
     "C0114",   # missing-module-docstring
     "C0115",   # missing-class-docstring
     "C0116",   # missing-function-docstring
+    "R0801",   # duplicate-code: false positive on short idiomatic patterns
     "R0902",   # too-many-instance-attributes: legitimate for model state classes
     "R0913",   # too-many-arguments: common in data science APIs
     "R0914",   # too-many-locals: acceptable in complex data processing functions

scripts/build_taxonomy.py

@@ -89,7 +89,10 @@ def build_rows(radii, osm, overture):
         .fillna("")
     )
     df = df.iloc[
-        sorted(range(len(df)), key = lambda i: (df.index[i].startswith("Other "), df.index[i]))
+        sorted(
+            range(len(df)),
+            key = lambda i: (df.index[i].startswith("Other "), df.index[i]),
+        )
     ]
     rows = []
     for label, row in df.iterrows():
@@ -302,7 +305,8 @@ def render(rows):
   </main>
 
   <footer>
-    OpenPOIs &mdash; by <a href="https://henryspatialanalysis.com/">Henry Spatial Analysis</a>
+    OpenPOIs &mdash;
+    by <a href="https://henryspatialanalysis.com/">Henry Spatial Analysis</a>
     &mdash; <a href="https://github.com/henryspatialanalysis/openpois">GitHub</a>
     &mdash; MIT License
   </footer>

scripts/check_taxonomy_sync.py

@@ -55,7 +55,11 @@ def parse_js_object_keys(text, name):
     )
     if not match:
         raise ValueError(f"Could not find `const {name}` in JS source")
-    return set(re.findall(r"^\s*([A-Za-z_][A-Za-z0-9_]*)\s*:", match.group(1), flags = re.MULTILINE))
+    return set(re.findall(
+        r"^\s*([A-Za-z_][A-Za-z0-9_]*)\s*:",
+        match.group(1),
+        flags = re.MULTILINE,
+    ))
 
 
 def report_diff(name, csv_set, js_set):

scripts/conflation/conflate.py

@@ -55,11 +55,6 @@
     find_spatial_candidates,
     select_best_matches,
 )
-
-CHECKPOINT_SUBDIR = "chunk_matches"
-DEDUP_CHECKPOINT_SUBDIR = "chunk_selfdedup"
-DEDUP_DROPPED_FILE = "overture_dedup_dropped.parquet"
-DEDUP_POST_FILTER_FILE = "overture_post_dedup.parquet"
 from openpois.conflation.merge import (
     build_merge_parts,
     build_merge_parts_chunked,
@@ -76,6 +71,11 @@
     load_top_level_matches,
 )
 
+CHECKPOINT_SUBDIR = "chunk_matches"
+DEDUP_CHECKPOINT_SUBDIR = "chunk_selfdedup"
+DEDUP_DROPPED_FILE = "overture_dedup_dropped.parquet"
+DEDUP_POST_FILTER_FILE = "overture_post_dedup.parquet"
+
 
 # -----------------------------------------------------------------
 # Memory instrumentation

scripts/exploratory/test_jax.py

@@ -52,11 +52,16 @@ def log_posterior(params, data):
     lp = 0.0
     lp += jnp.sum(stats.norm.logpdf(params['alpha'], loc = 0.0, scale = 5.0))
     lp += jnp.sum(stats.norm.logpdf(params['beta'], loc = 0.0, scale = 2.5))
-    lp += jnp.sum(stats.norm.logpdf(params['log_sigma'], loc = jnp.log(0.2), scale = 0.25))
+    lp += jnp.sum(stats.norm.logpdf(
+        params['log_sigma'], loc = jnp.log(0.2), scale = 0.25
+    ))
     mu = params['alpha'] + data['x'] @ params['beta']
-    lp += jnp.sum(stats.norm.logpdf(data['y'], loc = mu, scale = jnp.exp(params['log_sigma'])))
+    lp += jnp.sum(stats.norm.logpdf(
+        data['y'], loc = mu, scale = jnp.exp(params['log_sigma'])
+    ))
     return lp
 
+
 def predict(key, params, data, add_sigma: bool = True):
     mu = params['alpha'] + data['x'] @ params['beta']
     if add_sigma:

scripts/osm_data/data_viz.py

@@ -15,7 +15,8 @@
     directories.osm_data           — directory containing input parquet and viz/ output
     osm_data.tag_key               — tag key whose changes define observation
                                      events (used only in plot titles)
-    osm_data.timestamp_cols        — columns to parse as timestamps (rows with nulls dropped)
+    osm_data.timestamp_cols        — columns to parse as timestamps (rows with
+                                     nulls dropped)
     osm_data.top_n_types           — number of top shared labels in the multi-panel figure
     download.osm.end_date          — right-censoring date for still-unchanged tags
     osm_data.apply_model.model_stub — stub for loading model predictions
@@ -101,6 +102,7 @@ def fig_save(
         **kwargs
     )
 
+
 def get_preds_dict(model_stub: str | None) -> dict[str, pd.DataFrame]:
     """Load constant and shared-label model predictions."""
     if model_stub is None:

src/openpois/conflation/merge.py

@@ -716,7 +716,7 @@ def save_conflated_from_parts(
         p.unlink()
     part_paths[0].parent.rmdir()
 
-    print(f"  Done.")
+    print("  Done.")
     return n
 
 

src/openpois/models/jax_core.py

@@ -249,7 +249,10 @@ def nuts_sample_multichain(
             verbose = verbose,
         )
         # Add a leading singleton chain axis so downstream code is shape-uniform.
-        add_axis = lambda x: jnp.expand_dims(x, axis = 0)
+
+        def add_axis(x):
+            return jnp.expand_dims(x, axis = 0)
+
         return (
             jax.tree_util.tree_map(add_axis, draws),
             jax.tree_util.tree_map(add_axis, info),

tests/test_dedup_overture.py

@@ -4,7 +4,6 @@
 import geopandas as gpd
 import numpy as np
 import pandas as pd
-import pytest
 from shapely.geometry import Point
 
 from openpois.conflation.dedup_overture import (
@@ -141,7 +140,6 @@ def test_two_near_duplicates_make_one_pair(self):
         # Two POIs ~11 m apart with identical name + taxonomy →
         # strong match, emitted once.
         coords = [(-122.335, 47.608), (-122.335, 47.6081)]
-        n = 2
         shared_labels = np.array(["bank", "bank"], dtype = object)
         l0_bits = np.array([1, 1], dtype = np.uint16)
         names = np.array(["US Bank", "US Bank"], dtype = object)
@@ -175,12 +173,10 @@ def test_two_near_duplicates_make_one_pair(self):
 
     def test_far_points_no_pair(self):
         # Two POIs ~1 km apart → beyond 100 m max_radius.
-        coords = [(-122.335, 47.608), (-122.335, 47.618)]
         centroids = np.array(
             [[-122.335, 47.608], [-122.335, 47.618]],
             dtype = np.float64,
         )
-        n = 2
         pairs, _ = find_self_matches_chunked(
             centroids_lonlat = centroids,
             radii_m = np.array([100.0, 100.0], dtype = np.float32),

tests/test_format_observations.py

@@ -208,7 +208,8 @@ def test_keep_key_stickiness(self, tmp_path):
         amenities = list(out["amenity"].fillna(""))
         lasts = list(out["amenity_last_value"].fillna(""))
         assert amenities == ["restaurant", "restaurant", "bar"]
-        # v1: pre-change was None; v2: no change → last stays empty; v3: last = "restaurant".
+        # v1: pre-change was None; v2: no change → last stays empty;
+        # v3: last = "restaurant".
         assert lasts == ["", "", "restaurant"]
 
     def test_left_join_null_inheritance(self, tmp_path):