FrontierDevelopmentLab · mikeheyns · Mar 19, 2026 · Mar 18, 2026 · Mar 18, 2026 · Mar 18, 2026
diff --git a/README.md b/README.md
@@ -1,7 +1,33 @@
 # 2024-HL-GeoCL
 ## FDL-X Heliolab 2024: Geoeffectiveness Continuous Learning
 
-Currently two modules are included in the geocloak folder, namely DAGGER and SHEATH.
+GEO-CLoak (GEOeffectiveness Continual Learning Or Active Knowledge) is an end-to-end ML pipeline for forecasting geomagnetic perturbations. It combines two models to produce predictions at different lead times:
+
+- **SHEATH**: Predicts solar wind parameters (speed, density, temperature, magnetic field) from SDO solar imagery features. Uses a 3-layer MLP. Provides multi-day lead time, lower-fidelity forecasts.
+- **DAGGER-CL**: Predicts ground-level geomagnetic field perturbations (dBe, dBn, dBz) at ~175 SuperMAG magnetometer stations. Uses a GRU encoder with continual learning (Elastic Weight Consolidation) to adapt to new data without forgetting. Provides ~30-minute lead time, higher-fidelity forecasts.
+
+The pipeline flows: **Sun (SDO imagery) → SHEATH → solar wind forecast → DAGGER-CL (with real-time L1 data) → station predictions → global maps → web app**
+
+## Project Structure
+
+```
+geocloak/              Main Python package
+  sheath2024/          SHEATH model, dataloader, training
+  dagger-cl/           DAGGER-CL model, dataloaders, inference, continual learning
+  gp/                  Gaussian process & spherical harmonic interpolation
+  preprocess/          SDO image preprocessing (coronal hole/active region segmentation)
+  datautilus/           Real-time data download from NOAA SWPC
+
+public/                Standalone scripts for public use
+  inference_sheath.py  Run SHEATH inference from the command line
+  example_data/        Sample input files
+
+feature_vector_extraction/   Feature engineering pipeline (InfluxDB → 26 SHEATH features)
+updating_nrt_data/           Near-real-time data collection (ACE, DSCOVR, geomagnetic indices)
+app_dev/                     Streamlit web application
+scripts/                     Training scripts, utilities, and Jupyter notebooks
+models/                      Trained model checkpoints and scaler files
+```
 
 ## Setup
 

diff --git a/app_dev/streamlit_app/scripts/dagger.py b/app_dev/streamlit_app/scripts/dagger.py
@@ -93,8 +93,7 @@ def get_dagger():
     """
     formatted_timestamp = extract_timestamp_from_filename(last_uploaded_dbe_mean_graph)
 
-    st.html(
-        """
+    st.html("""
         <div class = "box">
             <div style="background-color: black; text-align: justify; border: 2px solid #ffffff; border-radius: 10px; box-shadow: 0 4px 8px 0 rgba(0, 0, 0, 0.2); transition: 0.3s; width: 100%; margin: auto; margin-top: 20px; padding: 20px; color: white;">
                 <div class="box-header">Geomagnetic Pertubations</div>
@@ -107,8 +106,7 @@ def get_dagger():
                 <p style="color:white; font-size:16px;">The yellow and red dots on the maps represent the ground stations or observational points where these measurements were taken. The color gradient on the map provides a visual representation of the magnetic field strength in nanoteslas (nT) for the respective components.</p>
 
             </div>
-        </div>"""
-    )
+        </div>""")
 
     # Create a 2x2 grid for images
     col1, col2 = st.columns(2)

diff --git a/geocloak/gp/gaussian_process_tutorial.ipynb b/geocloak/gp/gaussian_process_tutorial.ipynb
diff --git a/geocloak/gp/gaussian_process_tutorial_superMAG.ipynb b/geocloak/gp/gaussian_process_tutorial_superMAG.ipynb
diff --git a/geocloak/gp/spherical_elementary_currents_tutorial.ipynb b/geocloak/gp/spherical_elementary_currents_tutorial.ipynb
diff --git a/geocloak/gp/spherical_elementary_currents_tutorial_triangular.ipynb b/geocloak/gp/spherical_elementary_currents_tutorial_triangular.ipynb
diff --git a/geocloak/gp/spherical_harmonics.py b/geocloak/gp/spherical_harmonics.py
@@ -1,4 +1,4 @@
-""" Module for Spherical Harmonics
+"""Module for Spherical Harmonics
 
 Author: Opal Issan (PhD student @ UCSD). contact: oissan@ucsd.edu
 Last modified: July 21st, 2024

diff --git a/geocloak/gp/spherical_harmonics_tutorial.ipynb b/geocloak/gp/spherical_harmonics_tutorial.ipynb
diff --git a/geocloak/gp/supermag_api.py b/geocloak/gp/supermag_api.py
@@ -353,7 +353,7 @@ def SuperMAGGetInventory(logon, start, extent):
     urlstr = sm_coreurl("inventory.php", logon, start, extent)
 
     # get the string array of stations
-    (success, stations) = sm_GetUrl(urlstr, "raw")
+    success, stations = sm_GetUrl(urlstr, "raw")
 
     # if the inventory is valid extract the stations to an array
     # if an error occurs set the the ERROR keyword to be the error string
@@ -380,7 +380,7 @@ def SuperMAGGetIndices(logon, start, extent, flagstring="", **kwargs):
     urlstr += indices
 
     # get the string array of JSON data
-    (status, data_list) = sm_GetUrl(urlstr, "json")
+    status, data_list = sm_GetUrl(urlstr, "json")
 
     # default is to return a dataframe, but can also return an array
     if (kwargs.get("FORMAT", "none")).lower() == "list":
@@ -408,7 +408,7 @@ def SuperMAGGetData(logon, start, extent, flagstring, station, **kwargs):
     urlstr += indices
     urlstr += "&station=" + station.upper()
 
-    (status, data_list) = sm_GetUrl(urlstr, "json")
+    status, data_list = sm_GetUrl(urlstr, "json")
 
     # default is to return a dataframe, but can also return an array
     if (kwargs.get("FORMAT", "none")).lower() == "list":
@@ -437,12 +437,12 @@ def sm_microtest(choice, userid):
     start = [2019, 11, 15, 10, 40, 00]  # alt: start='2019-11-15T10:40'
 
     if choice == 1 or choice == 4:
-        (status, stations) = SuperMAGGetInventory(userid, start, 3600)
+        status, stations = SuperMAGGetInventory(userid, start, 3600)
         print(status)
         print(stations)
 
     if choice == 2 or choice == 4:
-        (status, data) = SuperMAGGetData(
+        status, data = SuperMAGGetData(
             userid, start, 3600, "all,delta=start,baseline=yearly", "HBK"
         )
         print(status)
@@ -465,7 +465,7 @@ def sm_microtest(choice, userid):
         plt.show()
 
     if choice == 3 or choice == 4:
-        (status, idxdata) = SuperMAGGetIndices(
+        status, idxdata = SuperMAGGetIndices(
             userid, start, 3600, "swiall,density,darkall,regall,smes"
         )
         # print(status)
@@ -486,41 +486,39 @@ def supermag_testing(userid):
 
     start = [2019, 11, 15, 10, 40, 00]  # alt: start='2019-11-15T10:40'
 
-    (status, stations) = SuperMAGGetInventory(userid, start, 3600)
+    status, stations = SuperMAGGetInventory(userid, start, 3600)
 
     # DATA fetches
     # BARE CALL, dataframe returned
-    (status, mydata1a) = SuperMAGGetData(userid, start, 3600, "", "HBK")
+    status, mydata1a = SuperMAGGetData(userid, start, 3600, "", "HBK")
     mydata1a  # is 1440 rows x 6 columns dataframe
     mydata1a.keys()  # Index(['tval', 'ext', 'iaga', 'N', 'E', 'Z'], dtype='object')
 
     # CALL with ALLINDICES, dataframe returned
-    (status, mydata1a) = SuperMAGGetData(userid, start, 3600, "all", "HBK")
+    status, mydata1a = SuperMAGGetData(userid, start, 3600, "all", "HBK")
     mydata1a  # is 1440 rows x 12 columns dataframe
     mydata1a.keys()  # Index(['tval', 'ext', 'iaga', 'glon', 'glat', 'mlt', 'mcolat', 'decl', 'sza', 'N', 'E', 'Z'], dtype='object')
 
     # BARE CALL, list returned
-    (status, mydata1b) = SuperMAGGetData(userid, start, 3600, "", "HBK", FORMAT="list")
+    status, mydata1b = SuperMAGGetData(userid, start, 3600, "", "HBK", FORMAT="list")
     len(mydata1b)  # is 1440 rows of dicts (key-value pairs)
     mydata1b[
         0:1
     ]  # {'tval': 1572726240.0, 'ext': 60.0, 'iaga': 'DOB', 'N': {'nez': -3.942651, 'geo': -5.964826}, 'E': {'nez': 4.492887, 'geo': 0.389075}, 'Z': {'nez': 7.608168, 'geo': 7.608168}}
 
     # CALL with ALLINDICES, list returned
-    (status, mydata1b) = SuperMAGGetData(
-        userid, start, 3600, "all", "HBK", FORMAT="list"
-    )
+    status, mydata1b = SuperMAGGetData(userid, start, 3600, "all", "HBK", FORMAT="list")
     mydata1b  # is 1440 rows of dicts (key-value pairs)
     mydata1b[
         0:1
     ]  # {'tval': 1572726240.0, 'ext': 60.0, 'iaga': 'DOB', 'glon': 9.11, 'glat': 62.07, 'mlt': 21.694675, 'mcolat': 30.361519, 'decl': 3.067929, 'sza': 124.698227, 'N': {'nez': -3.942651, 'geo': -5.964826}, 'E': {'nez': 4.492887, 'geo': 0.389075}, 'Z': {'nez': 7.608168, 'geo': 7.608168}}
 
     ####################
     # INDICES fetches
-    (status, idxdata) = SuperMAGGetIndices(userid, start, 3600)
+    status, idxdata = SuperMAGGetIndices(userid, start, 3600)
     idxdata  # empty!
 
-    (status, idxdata) = SuperMAGGetIndices(userid, start, 3600, "all,swiall,imfall")
+    status, idxdata = SuperMAGGetIndices(userid, start, 3600, "all,swiall,imfall")
     idxdata  # 1440 rows x 77 columns dataframe
     idxdata.keys()  # Index(['tval', 'SME', 'SML', 'SMLmlat', 'SMLmlt', 'SMLglat', 'SMLglon', 'SMLstid', 'SMU', 'SMUmlat', 'SMUmlt', 'SMUglat', 'SMUglon', 'SMUstid', 'SMEnum', 'SMEs', 'SMLs', 'SMLsmlat', 'SMLsmlt', 'SMLsglat', 'SMLsglon', 'SMLsstid', 'SMUs', 'SMUsmlat', 'SMUsmlt', 'SMUsglat', 'SMUsglon', 'SMUsstid', 'SMEsnum', 'SMEd', 'SMLd', 'SMLdmlat', 'SMLdmlt', 'SMLdglat', 'SMLdglon', 'SMLdstid', 'SMUd', 'SMUdmlat', 'SMUdmlt', 'SMUdglat', 'SMUdglon', 'SMUdstid', 'SMEdnum', 'SMEr', 'SMLr', 'SMLrmlat', 'SMLrmlt', 'SMLrglat', 'SMLrglon', 'SMLrstid', 'SMUr', 'SMUrmlat', 'SMUrmlt', 'SMUrglat', 'SMUrglon', 'SMUrstid', 'SMErnum', 'smr', 'smr00', 'smr06', 'smr12', 'smr18', 'smrnum', 'smrnum00', 'smrnum06', 'smrnum12', 'smrnum18', 'bgse', 'bgsm', 'vgse', 'vgsm', 'clockgse', 'clockgsm', 'density', 'dynpres', 'epsilon', 'newell'], dtype='object')
     #
@@ -561,7 +559,7 @@ def supermag_testing(userid):
     vgse_xyz = [(mydat["X"], mydat["Y"], mydat["Z"]) for mydat in vgse]  # grab all 3
 
     # We also offer a list format, for users who prefer to work in python lists
-    (status, mydata2c) = SuperMAGGetIndices(
+    status, mydata2c = SuperMAGGetIndices(
         userid, start, 3600, "all,swiall,imfall", FORMAT="list"
     )
     len(mydata2c)  # is 1440 rows of dicts (key-value pairs)