Update colab 2023 to fully fix #1011 and update some comments on notebook

Author: chriski777

jupyter/run_suite2p_colab_2023.ipynb

@@ -321,7 +321,7 @@
     "4. Classification of ROIs\n",
     "5. Spike Deconvolution\n",
     "\n",
-    "To run `registration`, `detection`, and `extraction` separately, we must first talk about a special class in `suite2p` called a `BinaryRWFile`. You can think of `BinaryRWFile` as a class for reading/writing image data that acts like a numpy array. "
+    "To run `registration`, `detection`, and `extraction` separately, we must first talk about a special class in `suite2p` called a `BinaryFile`. You can think of `BinaryFile` as a class for reading/writing image data that acts like a numpy array. "
    ]
   },
   {
@@ -355,7 +355,7 @@
     "id": "0idbURhWYsDM"
    },
    "source": [
-    "We'll run the registration module only on our example image which only contains data from a single channel. You can add in data for the second channel (e.g., `f_reg_chan2` and `f_raw_chan2`) using similar code to what we have above. Refer to the docs to see what the outputs refer to."
+    "We'll run the registration module only on our example image which only contains data from a single channel. You can add in data for the second channel (e.g., `f_reg_chan2` and `f_raw_chan2`) using similar code to what we have above. When writing a new `BinaryFile`, please make sure to specify the number of frames your `BinaryFile` instance will have. Refer to the docs to see what the outputs refer to."
    ]
   },
   {
@@ -382,7 +382,12 @@
    "source": [
     "### Running ROI Detection\n",
     "\n",
-    "To run ROI detection alone (called by the `detection_wrapper` function in the detection module), we'll first instantiate the necessary parameters. You only need a `BinaryRWFile` corresponding to a registered/unregistered recording. Here, we'll pass the `f_reg` we obtained after running the registration module above."
+    "To run ROI detection alone (called by the `detection_wrapper` function in the detection module), we'll first instantiate the necessary parameters. You only need a `BinaryRWFile` corresponding to a registered/unregistered recording. Here, we'll pass the `f_reg` we obtained after running the registration module above.\n",
+    "\n",
+    "Suite2p provides a default classification file containing a default dataset that is used to train a classifier that will be used for your data. One could specify their own classification file if they'd like. To do so, they should save a numpy file with a dict containing the following keys: \n",
+    "- `'stats'`: ROI Stats\n",
+    "- `'keys'`: keys of ROI stats that will be used for classification\n",
+    "- `'iscell'`: labels specifying whether an ROI is a cell or not\n"
    ]
   },
   {
@@ -394,7 +399,8 @@
    "outputs": [],
    "source": [
     "# Use default classification file provided by suite2p \n",
-    "classfile = suite2p.classification.builtin_classfile"
+    "classfile = suite2p.classification.builtin_classfile\n",
+    "np.load(classfile, allow_pickle=True)[()]"
    ]
   },
   {
@@ -416,7 +422,7 @@
    },
    "source": [
     "### Running Fluorescence Extraction\n",
-    "To run extraction alone (called by the `extraction_wrapper` function in the extraction module), we can just make use of any `stat` dictionary (from previous runs of suite2p or a custom user-made one). In this case, we'll use the one output by the cell above. If you'd like to extract signal, you can pass a `binaryRWFile` corresponding to the recording for the second channel to the `f_reg_chan2` parameter."
+    "To run extraction alone (called by the `extraction_wrapper` function in the extraction module), we can just make use of any `stat` dictionary (from previous runs of suite2p or a custom user-made one). In this case, we'll use the one output by the cell above. If you'd like to extract signal, you can pass a `binaryFile` corresponding to the recording for the second channel to the `f_reg_chan2` parameter."
    ]
   },
   {