Merge branch '68-tutorial-base-settings' into 69-tutorial-raw-readers

docs/tutorials/base_settings.ipynb

@@ -968,21 +968,20 @@
    "source": [
     "## MS data storage\n",
     "\n",
-    "The `spectrum_df` and `peak_df` will be saved into HDF5 (https://www.hdfgroup.org/solutions/hdf5/, https://docs.h5py.org/en/stable/) file by using readers' common API `save_hdf()`. The `spectrum_df` will be saved into `ms_data/spectrum_df`, and `peak_df` into `ms_data/peak_df`. We can use readers' `load_hdf()` method to load these two dataframes into the reader object. HDF5 files can be partially loaded or operated in memory-mapped mode for a RAM-limited environment, but this is not the topic of this tutorial.\n",
+    "The `spectrum_df` and `peak_df` will be saved into HDF5 (https://www.hdfgroup.org/solutions/hdf5/, https://docs.h5py.org/en/stable/) file by using readers' common API `save_hdf()`. The `spectrum_df` will be saved into `ms_data/spectrum_df`, and `peak_df` into `ms_data/peak_df`. We can use readers' `load_hdf()` method to load these two dataframes into the reader object.\n",
     "\n",
     "```\n",
     "# Save hdf\n",
     "ms_reader.save_hdf(hdf_path)\n",
     "\n",
     "# Load hdf\n",
     "ms_reader.load_hdf(hdf_path)\n",
-    "```"
+    "```\n",
+    "\n",
+    "If developers know how to use h5py package, it is also easy to directly load each column from the HDF5 file as a numpy array, and then re-construct `spectrum_df` and `peak_df` if necessary.\n",
+    "\n",
+    "HDF5 files can be partially loaded or operated in memory-mapped mode for a RAM-limited environment, but this is not the topic of this tutorial."
    ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": []
   }
  ],
  "metadata": {

tests/download_test_data.py

@@ -15,6 +15,7 @@ def download_file(url, local_filename):
                 f.write(chunk)
 
 
+raw_dir = "../nbs_tests/test_data"
 url_template = (
     "https://datashare.biochem.mpg.de/s/GOiZGGOhrHzS54M/download?path=%2F&files={}"
 )