[ENH] move transparent neurovault / NARPS from main repo

examples/plotting/plot_narps_transparency.py

@@ -0,0 +1,131 @@
+# %%
+# Comparing results across studies
+# --------------------------------
+#
+# NARPS study comparison
+#
+
+# %% Get data fromp neurovault
+from pathlib import Path
+
+from nilearn.datasets import fetch_neurovault_ids
+
+# {team : collections_id} dictionary
+collection_ids = {
+    "0JO0": "4807",
+    "6VV2": "4883",
+}
+
+# %%
+# Uncomment the code below to get images from all teams.
+# Note this will take a very long time.
+#
+# .. code-block:: python
+#
+#    import requests
+#    import pandas as pd
+#
+#    output_dir = Path.cwd() / "results" / "plot_transparency"
+#    output_dir.mkdir(exist_ok=True, parents=True)
+#
+#    output_file = output_dir / "analysis_pipelines.xlsx"
+#
+#
+#    def download_team_listing(output_file):
+#        url = 'https://github.com/poldrack/narps/raw/refs/heads/master/ImageAnalyses/metadata_files/analysis_pipelines_for_analysis.xlsx'
+#        response = requests.get(url, stream=True)
+#        with open(output_file, "wb") as file:
+#            for chunk in response.iter_content(chunk_size=8192):
+#                file.write(chunk)
+#
+#
+#    def get_collection_ids(infile=output_file):
+#        TEAMS_TO_SKIP = ['1K0E', 'X1Z4', 'L1A8', 'VG39', '5G9K', '16IN']
+#        # requires openpyxl
+#        teaminfo = pd.read_excel(infile, skiprows=1)
+#        collectionID = {}
+#        for t in teaminfo.index:
+#            teamID = teaminfo.loc[t, 'teamID']
+#            if teamID in TEAMS_TO_SKIP:
+#                continue
+#            public_link = teaminfo.loc[t, 'NV_collection_link']
+#            collectionID[teamID] = Path(public_link).stem
+#
+#        return(collectionID)
+#
+#    collection_ids = get_collection_ids()
+#
+
+nv_data = fetch_neurovault_ids(collection_ids=collection_ids.values())
+
+# %%
+# Select the data we want to plot.
+#
+# Keep only maps for hypothesis 1.
+#
+
+hypothesis = "1"
+
+data_to_plot = {k: {"collection_id": v} for k, v in collection_ids.items()}
+for x in nv_data["images_meta"]:
+    file = Path(x["file"])
+    if hypothesis not in file.stem:
+        continue
+
+    team = list(collection_ids.keys())[
+        list(collection_ids.values()).index(str(x["collection_id"]))
+    ]
+
+    if "_unthresh" in file.stem:
+        data_to_plot[team]["unthresholded"] = x["absolute_path"]
+    if "_thresh" in file.stem:
+        data_to_plot[team]["thresholded"] = x["absolute_path"]
+
+# %%
+# Compare the thresholded and unthresholded results
+#
+from matplotlib import pyplot as plt
+import numpy as np
+from nilearn.plotting import plot_stat_map, show
+
+threshold = 3
+figure_width = 9
+
+plotting_config = {
+    "display_mode": "x",
+    # take a cut through
+    # the ventro medial prefrontal cortex
+    # to match region of interest for hypothesis 1
+    "cut_coords": [1],
+    "draw_cross": False,
+    "vmax": 5,
+}
+
+
+fig, axes = plt.subplots(
+    len(data_to_plot),
+    2,
+    figsize=(figure_width, len(data_to_plot) * 4),
+)
+axes = np.atleast_2d(axes)
+
+for i, (team, data) in enumerate(data_to_plot.items()):
+    plot_stat_map(
+        data["thresholded"],
+        title=f"team {team} - thresholded",
+        axes=axes[i, 0],
+        cmap="inferno",
+        **plotting_config,
+    )
+    plot_stat_map(
+        data["unthresholded"],
+        title=f"team {team} - unthresholded",
+        transparency=data["unthresholded"],
+        transparency_range=[0, threshold],
+        axes=axes[i, 1],
+        cmap="cold_hot",
+        **plotting_config,
+    )
+
+show()
+