provide aerosol data via pooch (#267)

* provide aerosol data via pooch

* fix hash

* fix correct docstring

* changelog

CHANGELOG.rst

@@ -62,6 +62,11 @@ New Features
      <https://github.com/MESMER-group/mesmer/pull/220>`_). By `Mathias Hauser
      <https://github.com/mathause>`_.
 
+- The aerosol data is now automatically downloaded using `pooch <https://www.fatiando.org/pooch/latest/>`__.
+  (`#267 <https://github.com/MESMER-group/mesmer/pull/267>`_). By `Mathias Hauser
+  <https://github.com/mathause>`_.
+
+
 Breaking changes
 ^^^^^^^^^^^^^^^^
 

data/README.md

@@ -0,0 +1,3 @@
+# mesmer data
+
+This folder contains auxiliary data for mesmer. They are downloaded on demand using [pooch](https://www.fatiando.org/pooch/latest/).

environment.yml

@@ -12,6 +12,7 @@ dependencies:
  - numpy
  - packaging
  - pandas<2.0
+ - pooch
  - regionmask>=0.8
  - scikit-learn
  - sphinx

mesmer/calibrate_mesmer/train_gt.py

@@ -6,6 +6,7 @@
 Functions to train global trend module of MESMER.
 """
 
+import warnings
 
 import numpy as np
 import xarray as xr
@@ -126,7 +127,7 @@ def train_gt(var, targ, esm, time, cfg, save_params=True):
                 var_all = np.unique(var_all, axis=0)
 
             params_gt["saod"], params_gt["hist"] = train_gt_ic_OLSVOLC(
-                var_all, gt_lowess_hist, params_gt["time"]["hist"], cfg
+                var_all, gt_lowess_hist, params_gt["time"]["hist"]
             )
         elif params_gt["method"] == "LOWESS":
             params_gt["hist"] = gt_lowess_hist
@@ -201,7 +202,7 @@ def train_gt_ic_LOWESS(data):
     return gt_lowess, frac_lowess_name
 
 
-def train_gt_ic_OLSVOLC(var, gt_lowess, time, cfg):
+def train_gt_ic_OLSVOLC(var, gt_lowess, time, cfg=None):
     """
     Derive global trend (emissions + volcanoes) parameters from single ESM ic ensemble
     by adding volcanic spikes to LOWESS trend.
@@ -214,8 +215,8 @@ def train_gt_ic_OLSVOLC(var, gt_lowess, time, cfg):
         1d array of smooth global trend of variable
     time : np.ndarray
         1d array of years
-    cfg : module
-        config file containing metadata needed to load in stratospheric AOD time series
+    cfg : None
+        Passing cfg is no longer required.
 
     Returns
     -------
@@ -231,14 +232,17 @@ def train_gt_ic_OLSVOLC(var, gt_lowess, time, cfg):
 
     """
 
-    # specify necessary variables from cfg file
-    dir_obs = cfg.dir_obs
+    if cfg is not None:
+        warnings.warn(
+            "Passing ``cfg`` to ``train_gt_ic_OLSVOLC`` is no longer necessary",
+            FutureWarning,
+        )
 
     nr_runs, nr_ts = var.shape
 
     # account for volcanic eruptions in historical time period
     # load in observed stratospheric aerosol optical depth
-    aod_obs = load_strat_aod(time, dir_obs)
+    aod_obs = load_strat_aod(time)
     # drop "year" coords - aod_obs does not have coords (currently)
     aod_obs = aod_obs.drop_vars("year")
 

mesmer/core/_data.py

@@ -0,0 +1,25 @@
+import pooch
+
+import mesmer
+
+
+def fetch_remote_data(name):
+    """
+    uses pooch to cache files
+    """
+
+    cache_dir = pooch.os_cache("mesmer")
+
+    REMOTE_RESSOURCE = pooch.create(
+        path=cache_dir,
+        # The remote data is on Github
+        base_url="https://github.com/MESMER-group/mesmer/raw/{version}/data/",
+        registry={
+            "isaod_gl_2022.dat": "3d26e78bf0ee96a02c99e2a7a448dafda0ac847a5c914a75c7d9745e95fe68ee",
+        },
+        version=f"v{mesmer.__version__}",
+        version_dev="main",
+    )
+
+    # the file will be downloaded automatically the first time this is run.
+    return REMOTE_RESSOURCE.fetch(name)

mesmer/io/load_obs.py

@@ -7,11 +7,14 @@
 """
 
 import os
+import warnings
 
 import numpy as np
 import pandas as pd
 import xarray as xr
 
+from mesmer.core._data import fetch_remote_data
+
 
 def load_obs(targ, prod, lon, lat, cfg, sel_ref="native", ignore_nans=True):
     """Load observations which you previously downloaded.
@@ -163,15 +166,15 @@ def load_obs_tblend(prod, lon, lat, cfg, sel_ref):
     return tblend, time
 
 
-def load_strat_aod(time, dir_obs):
+def load_strat_aod(time, dir_obs=None):
     """Load observed global stratospheric aerosol optical depth time series.
 
     Parameters
     ----------
     time : np.ndarray
         1d array of years the AOD time series is required for
-    dir_obs : str
-        pathway to observations
+    dir_obs : None
+        Deprecated.
 
     Returns
     -------
@@ -184,9 +187,16 @@ def load_strat_aod(time, dir_obs):
       cimp6, 1850 - 2005 for cmip5)
     """
 
-    path_file = os.path.join(dir_obs, "aerosols", "isaod_gl.dat")
+    if dir_obs is not None:
+        warnings.warn(
+            "The aerosol data is now shipped with mesmer. Passing `dir_obs` to "
+            "``load_strat_aod`` is no longer necessary",
+            FutureWarning,
+        )
+
+    filename = fetch_remote_data("isaod_gl_2022.dat")
     df = pd.read_csv(
-        path_file,
+        filename,
         delim_whitespace=True,
         skiprows=11,
         names=("year", "month", "AOD"),

setup.cfg

@@ -36,6 +36,7 @@ install_requires =
     numpy
     packaging
     pandas < 2.0
+    pooch
     regionmask
     scikit-learn
     statsmodels