Fix: More fixing (#40)

emodel_generalisation/cli.py

@@ -108,7 +108,7 @@ def cli(verbose, no_progress):
 
 @cli.command("compute_currents")
 @click.option("--input-path", type=click.Path(exists=True), required=True)
-@click.option("--population_name", type=click.Path(exists=True), default=None)
+@click.option("--population-name", type=str, default=None)
 @click.option("--output-path", default="circuit_currents.h5", type=str)
 @click.option("--morphology-path", type=click.Path(exists=True), required=False)
 @click.option("--hoc-path", type=str, required=True)
@@ -184,30 +184,15 @@ def compute_currents(
         only_rin=only_rin,
     )
 
-    cols = ["resting_potential", "input_resistance", "exception"]
-    if not only_rin:
-        cols += ["holding_current", "threshold_current"]
-
-    # we populate the full circuit with duplicates if any
-    if len(cells_df) == len(unique_cells_df):
-        cells_df = unique_cells_df
-    else:
-        unique_cells_df = unique_cells_df.set_index(["morphology", "emodel"])
-        for entry, data in tqdm(
-            cells_df.groupby(["morphology", "emodel"]), disable=os.environ.get("NO_PROGRESS", False)
-        ):
-            for col in cols:
-                cells_df.loc[data.index, col] = unique_cells_df.loc[entry, col]
-
-    failed_cells = cells_df[
-        cells_df["input_resistance"].isna() | (cells_df["input_resistance"] < 0)
+    failed_cells = unique_cells_df[
+        unique_cells_df["input_resistance"].isna() | (unique_cells_df["input_resistance"] < 0)
     ].index
     if len(failed_cells) > 0:
         L.info("%s failed cells, we retry with fixed timesteps:", len(failed_cells))
         L.info(cells_df.loc[failed_cells])
         protocol_config["deterministic"] = False
-        cells_df.loc[failed_cells] = evaluate_currents(
-            cells_df.loc[failed_cells],
+        unique_cells_df.loc[failed_cells] = evaluate_currents(
+            unique_cells_df.loc[failed_cells],
             protocol_config,
             hoc_path,
             parallel_factory=parallel_factory,
@@ -216,13 +201,28 @@ def compute_currents(
             only_rin=only_rin,
         )
 
-        failed_cells = cells_df[
-            cells_df["input_resistance"].isna() | (cells_df["input_resistance"] < 0)
+        failed_cells = unique_cells_df[
+            unique_cells_df["input_resistance"].isna() | (unique_cells_df["input_resistance"] < 0)
         ].index
         if len(failed_cells) > 0:
             L.info("still %s failed cells (we drop):", len(failed_cells))
-            L.info(cells_df.loc[failed_cells])
-            cells_df.loc[failed_cells, "mtype"] = None
+            L.info(unique_cells_df.loc[failed_cells])
+            unique_cells_df.loc[failed_cells, "mtype"] = None
+
+    cols = ["resting_potential", "input_resistance", "exception"]
+    if not only_rin:
+        cols += ["holding_current", "threshold_current"]
+
+    # we populate the full circuit with duplicates if any
+    if len(cells_df) == len(unique_cells_df):
+        cells_df = unique_cells_df
+    else:
+        unique_cells_df = unique_cells_df.set_index(["morphology", "emodel"])
+        for entry, data in tqdm(
+            cells_df.groupby(["morphology", "emodel"]), disable=os.environ.get("NO_PROGRESS", False)
+        ):
+            for col in cols:
+                cells_df.loc[data.index, col] = unique_cells_df.loc[entry, col]
 
     cols_rename = {col: f"@dynamics:{col}" for col in cols if col != "exception"}
 
@@ -249,19 +249,10 @@ def plot_evaluation(cells_df, access_point, main_path="analysis_plot", clip=5, f
     """Make some plots of evaluations."""
     main_path = Path(main_path)
     main_path.mkdir(exist_ok=True)
-    if feature_filter is None or feature_filter == "":
-        feature_filter = FEATURE_FILTER
-        feature_filter.append("inv_time_to_first_spike")
-        feature_filter.append("burst_number")
-        feature_filter.append("ohmic_input_resistance_vb_ssse")
-        feature_filter.append("AHP_depth_abs")
-        feature_filter.append("AHP_depth")
-    else:
-        feature_filter = json.loads(feature_filter)
 
     L.info("Plotting summary figure...")
     scores = get_score_df(cells_df, filters=feature_filter)
-    cells_df["cost"] = np.clip(scores.max(1), 0, clip)
+    cells_df["cost"] = np.clip(scores.max(axis=1), 0, clip)
     _df = cells_df[["emodel", "mtype", "cost"]].groupby(["emodel", "mtype"]).mean().reset_index()
     plot_df = _df.pivot(index="emodel", columns="mtype", values="cost")
     plt.figure(figsize=(10, 6))
@@ -370,10 +361,6 @@ def evaluate(
         config_path, final_path, legacy, local_config=local_config_path
     )
     cells_df, _ = _load_circuit(input_path, morphology_path, population_name)
-    # cells_df = cells_df[cells_df.emodel == "bAC_L6BTC"]
-    # cells_df = cells_df[cells_df.mtype == "L23_LBC"].reset_index(drop=True)
-    # cells_df["@dynamics:AIS_scaler"] = 4.0
-    # cells_df["@dynamics:soma_scaler"] = 2.0
 
     if n_cells_per_emodel is not None:
         cells_df = (
@@ -443,6 +430,18 @@ def evaluate(
 
         exemplar_df = exemplar_df.set_index("emodel").loc[cells_df.emodel].reset_index()
 
+        if feature_filter is None or feature_filter == "":
+            feature_filter = FEATURE_FILTER
+            feature_filter.append("inv_time_to_first_spike")
+            feature_filter.append("burst_number")
+            feature_filter.append("ohmic_input_resistance_vb_ssse")
+            feature_filter.append("AHP_depth_abs")
+            feature_filter.append("AHP_depth")
+            feature_filter.append("RMPProtocol")
+            feature_filter.append("SearchHoldingCurrent")
+        else:
+            feature_filter = json.loads(feature_filter)
+
         pass_dfs = []
         Path(validation_path).mkdir(parents=True, exist_ok=True)
         with PdfPages(Path(validation_path) / "mm_features.pdf") as pdf:
@@ -458,7 +457,7 @@ def evaluate(
                     _pass[col] = cells_score_df[col] <= np.maximum(
                         5.0, 5.0 * exemplar_score_df[col].to_list()[0]
                     )
-                _pass = 1 - _pass.copy()  # copy to make it less fragmented
+                _pass = _pass.copy()  # copy to make it less fragmented
                 _pass["mtype"] = cells_df.mtype
 
                 data = _pass.groupby("mtype").mean()
@@ -471,7 +470,7 @@ def evaluate(
                     vmin=0.0,
                     vmax=1.0,
                     ax=ax,
-                    cbar_kws={"label": "Fraction of failed features", "shrink": 0.5},
+                    cbar_kws={"label": "Fraction of pass features", "shrink": 0.5},
                 )
 
                 ax.set_xlabel("mtype")

emodel_generalisation/version.py

@@ -19,4 +19,4 @@
 # or send a letter to Creative Commons, 171
 # Second Street, Suite 300, San Francisco, California, 94105, USA.
 
-VERSION = "0.2.7"  # pragma: no cover
+VERSION = "0.2.8.dev0"  # pragma: no cover

setup.py

@@ -50,6 +50,7 @@
     "sqlalchemy-utils>=0.37.2",
     "bluecellulab>=1.7.6",
     "voxcell>=3.1.5",
+    "efel>=5.5.5",
 ]
 
 doc_reqs = [