updating orientation distribution plots

src/roodmus/analysis/plot_alignment.py

@@ -161,9 +161,10 @@ def true_pose_distribution_plot(
     vmax: float | None = None,
 ):
     df_truth["euler_phi"] = df_truth["euler_phi"].astype(float)
-    df_truth["euler_theta"] = -(
-        df_truth["euler_theta"].astype(float) - np.pi / 2
-    )
+    df_truth["euler_theta"] = df_truth["euler_theta"].astype(float)
+    # df_truth["euler_theta"] = -(
+    #     df_truth["euler_theta"].astype(float) - np.pi / 2
+    # )
     df_truth["euler_psi"] = df_truth["euler_psi"].astype(float)
 
     grid = sns.jointplot(
@@ -194,7 +195,7 @@ def true_pose_distribution_plot(
     )
     grid.ax_joint.set_xticklabels(
         [
-            "\u03C0",
+            "-\u03C0",
             "-3/4\u03C0",
             "-\u03C0/2",
             "-\u03C0/4",
@@ -205,12 +206,12 @@ def true_pose_distribution_plot(
             "\u03C0",
         ]
     )
-    grid.ax_joint.set_yticks([-np.pi / 2, -np.pi / 4, 0, np.pi / 4, np.pi / 2])
+    grid.ax_joint.set_yticks([0, np.pi / 4, np.pi / 2, 3 / 4 * np.pi, np.pi])
     grid.ax_joint.set_yticklabels(
-        ["\u03C0/2", "\u03C0/4", "0", "\u03C0/4", "\u03C0/2"]
+        ["0", "\u03C0/4", "\u03C0/2", "3/4\u03C0", "\u03C0"]
     )
     grid.ax_joint.set_xlabel("Azimuth")
-    grid.ax_joint.set_ylabel("Elevation")
+    grid.ax_joint.set_ylabel("Tilt")
     # add new sublot to the right of the jointplot
     cbar_ax = grid.fig.add_axes([1, 0.15, 0.02, 0.7])
     # add colorbar to the new subplot
@@ -275,7 +276,11 @@ def make_and_save_plots(
                     "metadata_filename"
                 ].unique()
             ):
-                self.grid, self.vmin, self.vmax = picked_pose_distribution(
+                (
+                    self.grid,
+                    self.vmin,
+                    self.vmax,
+                ) = picked_pose_distribution_plot(
                     self.plot_data["plot_picked_pose_distribution"][
                         "df_picked"
                     ],
@@ -310,7 +315,7 @@ def _save_plot(self, meta_file: str):
             )
 
 
-def picked_pose_distribution(
+def picked_pose_distribution_plot(
     df_picked: pd.DataFrame,
     metadata_filename: str | List[str],
     vmin: float | None = None,
@@ -324,15 +329,15 @@ def picked_pose_distribution(
     )
 
     # change data type of column euler_phi to float
-    df_picked_grouped["euler_phi"] = df_picked_grouped["euler_phi"].astype(
-        float
-    )
-    df_picked_grouped["euler_theta"] = -(
-        df_picked_grouped["euler_theta"].astype(float) - np.pi / 2
-    )
-    df_picked_grouped["euler_psi"] = df_picked_grouped["euler_psi"].astype(
-        float
-    )
+    # df_picked_grouped["euler_phi"] = df_picked_grouped["euler_phi"].astype(
+    #     float
+    # )
+    # df_picked_grouped["euler_theta"] = -(
+    #     df_picked_grouped["euler_theta"].astype(float) - np.pi / 2
+    # )
+    # df_picked_grouped["euler_psi"] = df_picked_grouped["euler_psi"].astype(
+    #     float
+    # )
 
     # plot the alignment
     grid = sns.jointplot(
@@ -363,7 +368,7 @@ def picked_pose_distribution(
     )
     grid.ax_joint.set_xticklabels(
         [
-            "\u03C0",
+            "-\u03C0",
             "-3/4\u03C0",
             "-\u03C0/2",
             "-\u03C0/4",
@@ -374,12 +379,12 @@ def picked_pose_distribution(
             "\u03C0",
         ]
     )
-    grid.ax_joint.set_yticks([-np.pi / 2, -np.pi / 4, 0, np.pi / 4, np.pi / 2])
+    grid.ax_joint.set_yticks([0, np.pi / 4, np.pi / 2, 3 / 4 * np.pi, np.pi])
     grid.ax_joint.set_yticklabels(
-        ["-\u03C0/2", "-\u03C0/4", "0", "\u03C0/4", "\u03C0/2"]
+        ["0", "\u03C0/4", "\u03C0/2", "3/4\u03C0", "\u03C0"]
     )
     grid.ax_joint.set_xlabel("Azimuth")
-    grid.ax_joint.set_ylabel("Elevation")
+    grid.ax_joint.set_ylabel("Tilt")
     # add new sublot to the right of the jointplot
     cbar_ax = grid.fig.add_axes([1, 0.15, 0.02, 0.7])
     # add colorbar to the new subplot

src/roodmus/analysis/plot_picking.py

@@ -2215,13 +2215,13 @@ def plot_precision(
     sm = plt.cm.ScalarMappable(
         cmap="RdYlBu",
         norm=plt.Normalize(
-            vmin=df_precision["defocus"].min(),
-            vmax=df_precision["defocus"].max(),
+            vmin=df_precision["defocus"].min() / 10000,
+            vmax=df_precision["defocus"].max() / 10000,
         ),
     )
     sm._A = []
     cbar = fig.colorbar(sm, ax=ax)
-    cbar.set_label("defocus (Å)", rotation=270, labelpad=20, fontsize=12)
+    cbar.set_label("defocus (\u03bcm)", rotation=270, labelpad=20, fontsize=12)
     # add labels
     ax.set_xlabel("")
     ax.set_ylabel("precision", fontsize=14)
@@ -2283,13 +2283,13 @@ def plot_recall(
     sm = plt.cm.ScalarMappable(
         cmap="RdYlBu",
         norm=plt.Normalize(
-            vmin=df_precision["defocus"].min(),
-            vmax=df_precision["defocus"].max(),
+            vmin=df_precision["defocus"].min() / 10000,
+            vmax=df_precision["defocus"].max() / 10000,
         ),
     )
     sm._A = []
     cbar = fig.colorbar(sm, ax=ax)
-    cbar.set_label("defocus (Å)", rotation=270, labelpad=20, fontsize=12)
+    cbar.set_label("defocus (\u03bcm)", rotation=270, labelpad=20, fontsize=12)
     # add labels
     ax.set_xlabel("")
     ax.set_ylabel("recall", fontsize=14)

src/roodmus/analysis/utils.py

@@ -29,6 +29,7 @@
 import yaml
 import numpy as np
 from scipy.spatial import cKDTree
+from scipy.spatial.transform import Rotation as R
 from tqdm import tqdm
 import pandas as pd
 import pickle
@@ -59,26 +60,25 @@ def load_cs(self, cs_path):
 
     @classmethod
     def get_ugraph_cs(self, metadata_cs: np.recarray) -> List[str]:
-        """Grab micrograph file paths from .cs data.
+        """Grab micrograph file paths from .cs data. if micrograph file
+        paths are not present in the metadata, return an empty list.
 
         Args:
             metadata_cs (np.recarray): .cs file metadata.
 
         Returns:
             ugraph_paths (List[str]): micrograph file paths.
         """
-        ugraph_paths = metadata_cs["location/micrograph_path"].tolist()
 
-        # elif "blob/path" in metadata_cs.dtype.names:
-        #     ugraph_paths = metadata_cs["blob/path"]
-        # removed None from within the function to stop mypy errors
-        # else:
-        #     return None
+        if "location/micrograph_path" in metadata_cs.dtype.names:
+            ugraph_paths = metadata_cs["location/micrograph_path"].tolist()
+            ugraph_paths = [
+                os.path.basename(path).decode("utf-8").split("_")[-1]
+                for path in ugraph_paths
+            ]
+        else:
+            ugraph_paths = []
 
-        ugraph_paths = [
-            os.path.basename(path).decode("utf-8").split("_")[-1]
-            for path in ugraph_paths
-        ]
         return ugraph_paths
 
     @classmethod
@@ -164,13 +164,13 @@ def get_orientations_cs(self, metadata_cs: np.recarray, return_pose=False):
                 "alignments3D/pose"
             ]  # orientations as rodriques vectors
             # convert to euler angles
-            euler = np.array(
-                [geom.rot2euler(geom.expmap(p)) for p in pose],
-                dtype=float,
-            )
-            # euler = R.from_rotvec(pose).as_euler(
-            #     "zyx", degrees=False
-            # )  # convert to euler angles
+            # euler = np.array(
+            #     [geom.rot2euler(geom.expmap(p)) for p in pose],
+            #     dtype=float,
+            # )
+            euler = R.from_rotvec(pose).as_euler(
+                "ZYZ", degrees=False
+            )  # convert to euler angles
             if return_pose:
                 return euler, pose
             else:
@@ -205,7 +205,7 @@ def get_class2D_cs(self, metadata_cs: np.recarray):
             np.ndarray: 2D class.
         """
         if "alignments2D/class" in metadata_cs.dtype.names:
-            class2d = metadata_cs["alignments2D/class"]
+            class2d = metadata_cs["alignments2D/class"].astype(int)
         else:
             class2d = None
         return class2d
@@ -359,10 +359,16 @@ def get_orientations_star(self, metadata_star) -> np.ndarray:
         )
         if not euler_phi:  # if empty
             euler_phi = [np.nan] * num_particles
+        else:
+            euler_phi = [np.deg2rad(float(r)) for r in euler_phi]
         if not euler_theta:
             euler_theta = [np.nan] * num_particles
+        else:
+            euler_theta = [np.deg2rad(float(r)) for r in euler_theta]
         if not euler_psi:
             euler_psi = [np.nan] * num_particles
+        else:
+            euler_psi = [np.deg2rad(float(r)) for r in euler_psi]
         euler = np.stack([euler_phi, euler_theta, euler_psi], axis=1)
         return euler
 
@@ -376,7 +382,12 @@ def get_class2D_star(self, metadata_star):
         Returns:
             class2d (np.ndarray): 2D class data.
         """
-        class2d = metadata_star.column_as_list("particles", "_rlnClassNumber")
+        class2d = [
+            int(r)
+            for r in metadata_star.column_as_list(
+                "particles", "_rlnClassNumber"
+            )
+        ]
         if class2d:
             return np.array(class2d)
         else:
@@ -1017,7 +1028,7 @@ def _extract_from_metadata(
                 ugraph_filename = IO.get_ugraph_cs(m)
                 uid = IO.get_uid_cs(m)
                 tmp_results["uid"].extend(uid)
-                if ugraph_filename is None:
+                if not ugraph_filename:
                     num_particles = len(uid)
                     tmp_results["ugraph_filename"].extend(
                         [np.nan] * num_particles
@@ -1402,8 +1413,8 @@ def _extract_from_config(
                 position = instance["position"]
                 orientation = instance["orientation"]  # rotation vector
                 # convert to euler angles
-                euler = geom.rot2euler(geom.expmap(np.array(orientation)))
-                # euler = R.from_rotvec(orientation).as_euler("zyx")
+                # euler = geom.rot2euler(geom.expmap(np.array(orientation)))
+                euler = R.from_rotvec(orientation).as_euler("ZYZ")
                 if return_pose:
                     orientations_list.append(orientation)
                 else: