Eaa bs 1d

pyproject.toml

@@ -10,7 +10,6 @@ requires-python = ">=3.11"
 license = {file = "LICENSE.txt"}
 dependencies = [
     "sci-agent@git+https://github.com/mdw771/sci-agent",
-    "opencv-contrib-python",
     "matplotlib",
     "numpy",
     "scikit-image",

src/eaa/task_manager/tuning/analytical_focusing.py

@@ -43,6 +43,7 @@ def __init__(
         line_scan_tool_y_coordinate_args: Tuple[str, ...] = ("y_center",),
         image_acquisition_tool_x_coordinate_args: Tuple[str, ...] = ("x_center",),
         image_acquisition_tool_y_coordinate_args: Tuple[str, ...] = ("y_center",),
+        use_feature_tracking_subtask: bool = True,
         *args, **kwargs
     ):
         """Analytical scanning microscope focusing task manager driven
@@ -106,6 +107,9 @@ def __init__(
             See `line_scan_tool_x_coordinate_args`.
         image_acquisition_tool_y_coordinate_args: Tuple[str, ...]
             See `line_scan_tool_y_coordinate_args`.
+        use_feature_tracking_subtask: bool
+            If True, feature tracking will be run if the line scan feature has drifted
+            out of the FOV.
         """
         if acquisition_tool is None:
             raise ValueError("`acquisition_tool` must be provided.")
@@ -127,6 +131,7 @@ def __init__(
         self.last_acquisition_count_registered = 0
         self.last_acquisition_count_stitched = 0
 
+        self.use_feature_tracking_subtask: bool = use_feature_tracking_subtask
         self.feature_tracking_task_manager: Optional[AnalyticalFeatureTrackingTaskManager] = None
 
         self.line_scan_tool_x_coordinate_args = line_scan_tool_x_coordinate_args
@@ -168,7 +173,7 @@ def create_image_registration_tool(self, acquisition_tool: AcquireImage):
             reference_image=None,
             reference_pixel_size=1.0,
             image_coordinates_origin="top_left",
-            registration_method="sift",
+            registration_method="phase_correlation",
         )
         return image_registration_tool
 
@@ -292,6 +297,8 @@ def initialize_kwargs_buffers(
     ):
         self.line_scan_kwargs = copy.deepcopy(initial_line_scan_kwargs)
         self.image_acquisition_kwargs = copy.deepcopy(initial_2d_scan_kwargs)
+
+        self.line_scan_kwargs["view_scan_line_in_image"] = True
 
     def run_line_scan(self) -> float:
         """Run a line scan and return the FWHM of the Gaussian fit.
@@ -301,6 +308,7 @@ def run_line_scan(self) -> float:
         float
             The FWHM of the Gaussian fit.
         """
+        self.record_system_message(f"Acquiring line scan with {self.line_scan_kwargs}.")
         res = self.acquisition_tool.acquire_line_scan(**self.line_scan_kwargs)
         try:
             res = json.loads(res)
@@ -327,6 +335,7 @@ def update_optimization_model(self, fwhm: float):
         self.optimization_tool.update(x, -np.array([[fwhm]]))
 
     def run_2d_scan(self):
+        self.record_system_message(f"Acquiring 2D scan with {self.image_acquisition_kwargs}.")
         image_path = self.acquisition_tool.acquire_image(**self.image_acquisition_kwargs)
         content = f"Acquired 2D scan with kwargs: {self.image_acquisition_kwargs}"
         if isinstance(image_path, str):
@@ -457,16 +466,18 @@ def run_tuning_iteration(self, x: np.ndarray):
                 f"but got {len(x)} and {len(self.parameter_names)}."
             )
         x = np.array(x)
+        self.record_system_message(f"Setting parameters to {x}.")
         self.param_setting_tool.set_parameters(x)
         self.run_2d_scan()
         offset, is_present = self.find_offset_and_feature_presence()
-        if not is_present:
+        if not is_present and self.use_feature_tracking_subtask:
             msg = "Feature is not present in the current image. Running feature tracking sub-task."
             logger.info(msg)
             self.record_system_message(msg)
             offset = self.run_feature_tracking_subtask()
         if np.any(np.isnan(offset)):
             raise RuntimeError("Offset is NaN. Please set offset manually.")
+        self.record_system_message(f"Applying offset {offset}.")
         self.apply_offset_to_kwargs_buffers(offset)
         fwhm = self.run_line_scan()
         if np.isnan(fwhm):

src/eaa/tool/imaging/aps_mic/acquisition.py

@@ -55,6 +55,7 @@ def __init__(
         show_colorbar_in_image: bool = False,
         require_approval: bool = False,
         line_scan_return_gaussian_fit: bool = False,
+        scan_samy: bool = False,
         *args, **kwargs
     ):
         """Image acquisition tool with Bluesky.
@@ -86,7 +87,9 @@ def __init__(
         line_scan_return_gaussian_fit: bool, optional
             If True, the function returns a stringified JSON object containing the image path
             and the Gaussian fit FWHM.
-
+        scan_samy: bool, optional
+            If True, the line_scan is generated by moving sample-y motor
+
         Raises
         ------
         ImportError
@@ -106,6 +109,7 @@ def __init__(
         self.plot_image_in_log_scale = plot_image_in_log_scale
         self.show_colorbar_in_image = show_colorbar_in_image
         self.line_scan_return_gaussian_fit = line_scan_return_gaussian_fit
+        self.scan_samy = scan_samy
         super().__init__(*args, require_approval=require_approval, **kwargs)
 
 
@@ -294,7 +298,7 @@ def acquire_line_scan(
 
                 img_path, [_, _, _, fwhm] = save_xrf_line_scan(
                     mda_file_path, png_output_dir, roi_num=self.xrf_roi_num, 
-                    return_line_array=True
+                    return_line_array=True, scan_samy=self.scan_samy,
                 )
                 wait_for_file(img_path, duration=5)
 

src/eaa/tool/imaging/aps_mic/param_tuning.py

@@ -86,6 +86,7 @@ def set_parameters(
             self.RE(self.bps.mv(self.zp_z_motor, parameters[0]))
             msg = f"Moved Zone Plate z position to position: {parameters[0]}"
             logger.info(msg)
+            self.update_parameter_history(parameters)
             return msg
         else:
             raise ValueError("parameter_ranges is not set")

src/eaa/tool/imaging/aps_mic/scan_control.py

@@ -35,6 +35,7 @@ def __init__(self, require_approval: bool = False, *args, **kwargs):
         self.acquire_image_tool.scan2d_plan = fly2d_scanrecord
         self.acquire_image_tool.scan1d_plan = step1d_scanrecord
         self.acquire_image_tool.samy_motor = oregistry["samy"]
+        self.acquire_image_tool.scan_samy = True
 
         self.param_tuning_tool = BlueskySetParameters()
         self.param_tuning_tool.RE = RE

src/eaa/tool/imaging/aps_mic/util.py

@@ -147,7 +147,7 @@ def process_xrfdata(
         return None
 
 
-def plot_xrf_line_scan(x, y, val_gauss, fwhm, scan_name, roi_num):
+def plot_xrf_line_scan(x, y, val_gauss, fwhm, scan_name, roi_num, scan_samy=False):
     """
     Plot the XRF line scan data.
     """
@@ -165,7 +165,7 @@ def plot_xrf_line_scan(x, y, val_gauss, fwhm, scan_name, roi_num):
     )
 
     ax.legend()
-    ax.set_xlabel("X-axis Position")
+    ax.set_xlabel("X-axis Position" if not scan_samy else "Y-axis Position")
     ax.set_ylabel("Intensity")
     ax.set_title(f"{scan_name}-{roi_num}")
     ax.grid(True)
@@ -178,7 +178,8 @@ def save_xrf_line_scan(
     output_dir: str, 
     roi_num: int,
     y_threshold: float = 0.0,
-    return_line_array: bool = False
+    return_line_array: bool = False,
+    scan_samy: bool = False,
 ) -> str | None:
 
     """
@@ -196,6 +197,8 @@ def save_xrf_line_scan(
         The threshold for the Gaussian fit.
     return_line_array : bool
         If True, the line array will be returned.
+    scan_samy : bool
+        If True, line profile is generated using sample-y motor, scanning sample-y
 
     Returns
     -------
@@ -233,7 +236,7 @@ def save_xrf_line_scan(
 
         # Plot the data and the fit
         scan_name = os.path.basename(mda_path)
-        fig = plot_xrf_line_scan(x, y, val_gauss, fwhm, scan_name, roi_num)
+        fig = plot_xrf_line_scan(x, y, val_gauss, fwhm, scan_name, roi_num, scan_samy=scan_samy)
         sc = scan_name.replace(".mda", "")
         fname = f"{output_dir}/{sc}_ROI{roi_num}.png"
         fig.savefig(fname)

src/eaa/tool/imaging/registration.py

@@ -2,9 +2,9 @@
 import logging
 import json
 
-import cv2
 import numpy as np
 import scipy.ndimage as ndi
+from skimage import feature
 from sciagent.tool.base import BaseTool, check, ToolReturnType, tool
 
 from eaa.tool.imaging.acquisition import AcquireImage
@@ -261,7 +261,7 @@ def prepare_image_for_feature_matching(self, image: np.ndarray) -> np.ndarray:
             image = (image - min_val) / (max_val - min_val)
         else:
             image = np.zeros_like(image, dtype=np.float32)
-        return (image * 255).astype(np.uint8)
+        return image
 
     def adjust_points_for_origin(
         self,
@@ -282,30 +282,39 @@ def feature_based_registration(
         image_t: np.ndarray,
         image_r: np.ndarray,
     ) -> np.ndarray:
-        image_t_uint8 = self.prepare_image_for_feature_matching(image_t)
-        image_r_uint8 = self.prepare_image_for_feature_matching(image_r)
-        sift = cv2.SIFT_create()
-        keypoints_t, descriptors_t = sift.detectAndCompute(image_t_uint8, None)
-        keypoints_r, descriptors_r = sift.detectAndCompute(image_r_uint8, None)
+        image_t_float = self.prepare_image_for_feature_matching(image_t)
+        image_r_float = self.prepare_image_for_feature_matching(image_r)
+        sift_t = feature.SIFT()
+        sift_r = feature.SIFT()
+        sift_t.detect_and_extract(image_t_float)
+        sift_r.detect_and_extract(image_r_float)
+
+        descriptors_t = sift_t.descriptors
+        descriptors_r = sift_r.descriptors
+        keypoints_t = sift_t.keypoints
+        keypoints_r = sift_r.keypoints
 
-        if descriptors_t is None or descriptors_r is None:
+        if (
+            descriptors_t is None
+            or descriptors_r is None
+            or descriptors_t.size == 0
+            or descriptors_r.size == 0
+        ):
             raise RuntimeError("SIFT feature detection failed to find descriptors.")
 
-        matcher = cv2.BFMatcher(cv2.NORM_L2)
-        raw_matches = matcher.knnMatch(descriptors_t, descriptors_r, k=2)
-        good_matches = []
-        for match_pair in raw_matches:
-            if len(match_pair) != 2:
-                continue
-            m, n = match_pair
-            if m.distance < 0.75 * n.distance:
-                good_matches.append(m)
+        matches = feature.match_descriptors(
+            descriptors_t,
+            descriptors_r,
+            metric="euclidean",
+            cross_check=True,
+            max_ratio=0.75,
+        )
 
-        if len(good_matches) < 3:
+        if matches.shape[0] < 3:
             raise RuntimeError("Not enough SIFT matches to estimate translation.")
 
-        pts_t = np.array([keypoints_t[m.queryIdx].pt for m in good_matches])
-        pts_r = np.array([keypoints_r[m.trainIdx].pt for m in good_matches])
+        pts_t = keypoints_t[matches[:, 0]][:, ::-1]
+        pts_r = keypoints_r[matches[:, 1]][:, ::-1]
         pts_t = self.adjust_points_for_origin(pts_t, image_t.shape)
         pts_r = self.adjust_points_for_origin(pts_r, image_r.shape)
         deltas = pts_r - pts_t