fix RUFF rules

pyfixest/did/did2s.py

@@ -232,7 +232,7 @@ def _did2s_estimate(
     _first_u = data[f"{yname}"].to_numpy().flatten() - Y_hat
     data[f"{yname}_hat"] = _first_u
 
-    # intercept needs to be dropped by hand due to the presence of fixed effects in the first stage  # noqa: W505
+    # intercept needs to be dropped by hand due to the presence of fixed effects in the first stage
     fit2 = cast(
         Feols,
         feols(
@@ -302,7 +302,7 @@ def _did2s_vcov(
     else:
         weights_array = np.sqrt(data[weights].to_numpy())
 
-    # some formula parsing to get the correct formula for the first and second stage model matrix  # noqa: W505
+    # some formula parsing to get the correct formula for the first and second stage model matrix
     first_stage_x, first_stage_fe = first_stage.split("|")
     first_stage_fe_list = [f"C({i})" for i in first_stage_fe.split("+")]
     first_stage_fe_fml = "+".join(first_stage_fe_list)

pyfixest/estimation/FixestMulti_.py

@@ -361,7 +361,7 @@ def _estimate_all_models(
                         FIT.wls_transform()
 
                     FIT.get_fit()
-                    # if X is empty: no inference (empty X only as shorthand for demeaning)  # noqa: W505
+                    # if X is empty: no inference (empty X only as shorthand for demeaning)
                     if not FIT._X_is_empty:
                         # inference
                         vcov_type = _get_vcov_type(vcov, fval)

pyfixest/estimation/FormulaParser.py

@@ -746,7 +746,7 @@ def _dict_to_list_of_formulas(unpacked: dict[str, list[str]]) -> list[str]:
         else:
             variable_fml = [res["variable"][i] for i in range(len(res["variable"]))]
         if variable_type in ["sw0", "csw0"]:
-            variable_fml = ["0"] + variable_fml
+            variable_fml = ["0", *variable_fml]
 
     fml_list = []
     if variable_fml:

pyfixest/estimation/feols_.py

@@ -738,7 +738,7 @@ def _vcov_crv3_fast(self, clustid, cluster_col):
         tXX = np.transpose(_X) @ _X
         tXy = np.transpose(_X) @ _Y
 
-        # compute leave-one-out regression coefficients (aka clusterjacks')  # noqa: W505
+        # compute leave-one-out regression coefficients (aka clusterjacks')
         for ixg, g in enumerate(clustid):
             Xg = _X[np.equal(g, cluster_col)]
             Yg = _Y[np.equal(g, cluster_col)]
@@ -2420,14 +2420,14 @@ def _check_vcov_input(vcov: Union[str, dict[str, str]], data: pd.DataFrame):
     """
     assert isinstance(vcov, (dict, str, list)), "vcov must be a dict, string or list"
     if isinstance(vcov, dict):
-        assert list(vcov.keys())[0] in [
+        assert next(iter(vcov.keys())) in [
             "CRV1",
             "CRV3",
         ], "vcov dict key must be CRV1 or CRV3"
         assert isinstance(
-            list(vcov.values())[0], str
+            next(iter(vcov.values())), str
         ), "vcov dict value must be a string"
-        deparse_vcov = list(vcov.values())[0].split("+")
+        deparse_vcov = next(iter(vcov.values())).split("+")
         assert len(deparse_vcov) <= 2, "not more than twoway clustering is supported"
 
     if isinstance(vcov, list):
@@ -2471,8 +2471,8 @@ def _deparse_vcov_input(vcov: Union[str, dict[str, str]], has_fixef: bool, is_iv
         The name of the cluster variable.
     """
     if isinstance(vcov, dict):
-        vcov_type_detail = list(vcov.keys())[0]
-        deparse_vcov = list(vcov.values())[0].split("+")
+        vcov_type_detail = next(iter(vcov.keys()))
+        deparse_vcov = next(iter(vcov.values())).split("+")
         if isinstance(deparse_vcov, str):
             deparse_vcov = [deparse_vcov]
         deparse_vcov = [x.replace(" ", "") for x in deparse_vcov]

pyfixest/estimation/feols_compressed_.py

@@ -164,7 +164,7 @@ def prepare_model_matrix(self):
                     + [col for col in data_long_mundlak.columns if col != "Intercept"]
                 )
 
-                self._coefnames = ["Intercept"] + covars_updated
+                self._coefnames = ["Intercept", *covars_updated]
                 self._fe = None
 
         else:

pyfixest/estimation/fepois_.py

@@ -459,7 +459,7 @@ def _check_for_separation(
 
     if separation_na:
         warnings.warn(
-            f"{str(len(separation_na))} observations removed because of separation."
+            f"{len(separation_na)!s} observations removed because of separation."
         )
 
     return list(separation_na)

pyfixest/report/summarize.py

@@ -117,7 +117,7 @@ def etable(
     pandas.DataFrame
         A styled DataFrame with the coefficients and standard errors of the models.
         When output is "tex", the LaTeX code is returned as a string.
-    """  # noqa: D301
+    """
     if signif_code is None:
         signif_code = [0.001, 0.01, 0.05]
     assert (
@@ -689,7 +689,8 @@ def _parse_coef_fmt(coef_fmt: str, custom_stats: dict):
         r"\[",
         r"\]",
         ",",
-    ] + custom_elements
+        *custom_elements,
+    ]
     allowed_elements.sort(key=len, reverse=True)
 
     coef_fmt_elements = re.findall("|".join(allowed_elements), coef_fmt)
@@ -1249,7 +1250,7 @@ def mean_newline_std(x):
 
     # Calculate the desired statistics
     if (byrow is not None) and (bycol is not None):
-        bylist = [byrow] + bycol
+        bylist = [byrow, *bycol]
         res = df.groupby(bylist).agg(agg_funcs)
     if (byrow is None) and (bycol is None):
         res = df.agg(agg_funcs)
@@ -1342,7 +1343,7 @@ def mean_newline_std(x):
             # Finally we want to have the objects first and then the statistics
             if not isinstance(res.columns, pd.MultiIndex):
                 res.columns = pd.MultiIndex.from_tuples(res.columns)
-            res.columns = res.columns.reorder_levels(bycol + ["Statistics"])
+            res.columns = res.columns.reorder_levels([*bycol, "Statistics"])
             # And sort it properly by the variables
             # (we want to preserve the order of the lowest level for the stats)
             levels_to_sort = list(range(res.columns.nlevels - 1))

pyproject.toml

@@ -135,6 +135,7 @@ select = [
   "D",   # flake8-docstrings
   "UP",  # pyupgrade
   "SIM", # flake8-simplify
+  "RUF", # ruff,
   "TRY", # tryceratops
 ]
 

tests/test_did.py

@@ -176,7 +176,7 @@ def test_errors(data):
     rng = np.random.default_rng(12)
     data["treat2"] = rng.choice([0, 1, 2], size=len(data))
     with pytest.raises(ValueError):
-        fit = did2s_pyfixest(  # noqa: F841
+        fit = did2s_pyfixest(
             data,
             yname="dep_var",
             first_stage="~ X | state + year",

tests/test_poisson.py

@@ -35,7 +35,7 @@ def test_separation():
     with pytest.warns(
         UserWarning, match="2 observations removed because of separation."
     ):
-        fepois("Y ~ X  | fe1", data=example1, vcov="hetero", separation_check=["fe"])  # noqa: F841
+        fepois("Y ~ X  | fe1", data=example1, vcov="hetero", separation_check=["fe"])
 
     if False:
         # this example is taken from ppmlhdfe's primer on separation https://github.com/sergiocorreia/ppmlhdfe/blob/master/guides/separation_primer.md
@@ -52,7 +52,7 @@ def test_separation():
         with pytest.warns(
             UserWarning, match="2 observations removed because of separation."
         ):
-            fepois("Y ~ X1 + X2", data=example2, vcov="hetero", separation_check=["ir"])  # noqa: F841
+            fepois("Y ~ X1 + X2", data=example2, vcov="hetero", separation_check=["ir"])
 
     # ppmlhdfe test data sets (check readme in data/ppmlhdfe_separation_examples)
     path = os.path.dirname(os.path.abspath(__file__))

tests/test_predict_resid_fixef.py

@@ -149,7 +149,7 @@ def test_vs_fixest(data, fml):
         raise ValueError("Predictions for OLS are not of the same length.")
 
     # test predict for Poisson
-    # if not np.allclose(fepois_mod.predict(data = data2), stats.predict(r_fixest_pois, newdata = data2)):  # noqa: W505
+    # if not np.allclose(fepois_mod.predict(data = data2), stats.predict(r_fixest_pois, newdata = data2)):
     #    raise ValueError("Predictions for Poisson are not equal")
 
     # test resid for OLS

tests/test_ses.py

@@ -205,7 +205,7 @@ def test_CRV3_fixef(N, seed, beta_type, error_type):
 @pytest.mark.slow
 def run_crv3_poisson():
     data = get_data(N=1000, seed=1234, beta_type="1", error_type="1", model="Fepois")
-    fit = fepois(  # noqa: F841
+    fit = fepois(
         fml="Y~X1 + C(f2)",
         data=data,
         vcov={"CRV3": "f1"},

tests/test_vs_fixest.py

@@ -50,14 +50,14 @@
     ("Y ~ X1 + i(f1, ref = 1) | f2 + f3"),
     ("Y ~ X1 + i(f1) + i(f2)"),
     ("Y ~ X1 + i(f1, ref = 1) + i(f2, ref = 2)"),
-    # ("Y ~ X1 + C(f1):C(fe2)"),                  # currently does not work as C():C() translation not implemented # noqa: W505
-    # ("Y ~ X1 + C(f1):C(fe2) | f3"),             # currently does not work as C():C() translation not implemented # noqa: W505
+    # ("Y ~ X1 + C(f1):C(fe2)"),                  # currently does not work as C():C() translation not implemented
+    # ("Y ~ X1 + C(f1):C(fe2) | f3"),             # currently does not work as C():C() translation not implemented
     ("Y ~ X1 + X2:f1"),
     ("Y ~ X1 + X2:f1 | f3"),
     ("Y ~ X1 + X2:f1 | f3 + f1"),
-    # ("log(Y) ~ X1:X2 | f3 + f1"),               # currently, causes big problems for Fepois (takes a long time) # noqa: W505
-    # ("log(Y) ~ log(X1):X2 | f3 + f1"),          # currently, causes big problems for Fepois (takes a long time) # noqa: W505
-    # ("Y ~  X2 + exp(X1) | f3 + f1"),            # currently, causes big problems for Fepois (takes a long time) # noqa: W505
+    # ("log(Y) ~ X1:X2 | f3 + f1"),               # currently, causes big problems for Fepois (takes a long time)
+    # ("log(Y) ~ log(X1):X2 | f3 + f1"),          # currently, causes big problems for Fepois (takes a long time)
+    # ("Y ~  X2 + exp(X1) | f3 + f1"),            # currently, causes big problems for Fepois (takes a long time)
     ("Y ~ X1 + i(f1,X2)"),
     ("Y ~ X1 + i(f1,X2) + i(f2, X2)"),
     ("Y ~ X1 + i(f1,X2, ref =1) + i(f2)"),
@@ -69,14 +69,14 @@
     ("Y ~ X1 + i(f2,X2, ref=2.0)"),
     ("Y ~ X1 + i(f1,X2, ref=3.0) | f2"),
     ("Y ~ X1 + i(f1,X2, ref=4.0) | f2 + f3"),
-    # ("Y ~ C(f1):X2"),                          # currently does not work as C():X translation not implemented # noqa: W505
+    # ("Y ~ C(f1):X2"),                          # currently does not work as C():X translation not implemented
     # ("Y ~ C(f1):C(f2)"),                       # currently does not work
     ("Y ~ X1 + I(X2 ** 2)"),
     ("Y ~ X1 + I(X1 ** 2) + I(X2**4)"),
     ("Y ~ X1*X2"),
     ("Y ~ X1*X2 | f1+f2"),
-    # ("Y ~ X1/X2"),                             # currently does not work as X1/X2 translation not implemented # noqa: W505
-    # ("Y ~ X1/X2 | f1+f2"),                     # currently does not work as X1/X2 translation not implemented # noqa: W505
+    # ("Y ~ X1/X2"),                             # currently does not work as X1/X2 translation not implemented
+    # ("Y ~ X1/X2 | f1+f2"),                     # currently does not work as X1/X2 translation not implemented
     ("Y ~ X1 + poly(X2, 2) | f1"),
 ]
 
@@ -470,10 +470,10 @@ def test_single_fit_fepois(
     check_absolute_diff(py_deviance, r_deviance, 1e-08, "py_deviance != r_deviance")
 
     if not mod._has_fixef:
-        py_predict_response = mod.predict(type="response")  # noqa: F841
-        py_predict_link = mod.predict(type="link")  # noqa: F841
-        r_predict_response = stats.predict(r_fixest, type="response")  # noqa: F841
-        r_predict_link = stats.predict(r_fixest, type="link")  # noqa: F841
+        py_predict_response = mod.predict(type="response")
+        py_predict_link = mod.predict(type="link")
+        r_predict_response = stats.predict(r_fixest, type="response")
+        r_predict_link = stats.predict(r_fixest, type="link")
         check_absolute_diff(
             py_predict_response[0:5],
             r_predict_response[0:5],