Merge pull request #91 from oubino/new_working

New working
oubino · Feb 8, 2024 · 6c26a22 · 6c26a22
2 parents 12c37b1 + 0b3c258
commit 6c26a22
Show file tree

Hide file tree

Showing 30 changed files with 180 additions and 3,295 deletions.
diff --git a/README.rst b/README.rst
@@ -16,8 +16,7 @@ For more comprehensive documentation please see https://oubino.github.io/locpix/
 #. Segmentation:
 
    #. `Classic segmentation`_ : Use classic method to segment histograms to extract relevant localisations
-   #. `Cellpose segmentation (no training)`_ : Use Cellpose method to segment histograms to extract relevant localisations with no retraining of Cellpose model
-   #. `Cellpose segmentation (training)`_ : Use Cellpose method to segment histograms to extract relevant localisations with retraining of Cellpose model
+   #. `Cellpose segmentation`_ : Use Cellpose method to segment histograms to extract relevant localisations
    #. `Ilastik segmentation`_ : Use Ilastik method to segment histograms to extract relevant localisations
 
 #. `Membrane performance`_ : Performance metrics calculation based on the localisations (not the histograms!)
@@ -36,12 +35,10 @@ will not be accessed again!
 Usage configuration
 -------------------
 
-Each script can be run with a GUI, but can also be run in headless mode.
-
-In headless mode each script needs a configuration file (.yaml file), which should be
+Each script needs a configuration file (.yaml file), which should be
 specified using the -c flag.
 
-Each configuration used, whether run in GUI or headless mode will be saved in the project directory.
+Each configuration used will be saved in the project directory.
 
 The templates for the configuration files can be found in the `templates folder <https://github.com/oubino/locpix/tree/master/src/locpix/templates>`_.
 
@@ -81,13 +78,7 @@ This script preprocesses the input .csv data for later use AND **must be run fir
 This script will take in .csv files, and convert them to .parquet files,
 while also wrangling the data into our data format.
 
-To run the script using the GUI, run
-
-.. code-block:: console
-
-   (locpix-env) $ preprocess
-
-To run the script without a GUI -i -c and -o flags should be specified
+To run the script -i -c and -o flags should be specified
 
 .. code-block:: console
 
@@ -98,13 +89,7 @@ Annotate
 
 This script allows for manual segmentation of the localisations.
 
-To run the script using the GUI, run
-
-.. code-block:: console
-
-   (locpix-env) $ annotate
-
-To run the script without a GUI -i and -c flags should be specified
+To run the script -i and -c flags should be specified
 
 .. code-block:: console
 
@@ -118,13 +103,7 @@ Get markers
 
 This script allows for labelling the localisation image with a marker to represent the cells.
 
-To run the script using the GUI, run
-
-.. code-block:: console
-
-   (locpix-env) $ get_markers
-
-To run the script without a GUI -i and -c flags should be specified
+To run the script -i and -c flags should be specified
 
 .. code-block:: console
 
@@ -135,106 +114,59 @@ Classic segmentation
 
 Perform classic segmentation on our localisation dataset.
 
-To run the script using the GUI, run
-
-.. code-block:: console
-
-   (locpix-env) $ classic
-
-To run the script without a GUI -i and -c flags should be specified
+To run the script -i and -c flags should be specified
 
 .. code-block:: console
 
    (locpix-env) $ classic -i path/to/project/directory -c path/to/config/file
 
-Cellpose segmentation (no training)
-^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-
-   Need to activate extra requirements - these are big and not included in initial install.
-
-   Note that if you have a GPU this will speed this up.
+Cellpose segmentation
+^^^^^^^^^^^^^^^^^^^^^
 
-   Note we modified Cellpose to fit in with our analysis, therefore you need to install our forked repository - note below will clone the Cellpose repository to wherever you are located
+Need to activate extra requirements - these are big and not included in initial install.
 
-   If you have a GPU
-
-   .. code-block:: console
+Note that if you have a GPU this will speed this up.
 
-      (locpix-env) $ pip install torch torchvision --extra-index-url https://download.pytorch.org/whl/cu117
-      (locpix-env) $ git clone https://github.com/oubino/cellpose
-      (locpix-env) $ cd cellpose
-      (locpix-env) $ pip install .
+Note we modified Cellpose to fit in with our analysis, therefore you need to install our forked repository - note below will clone the Cellpose repository to wherever you are located
 
-   If you don't have a GPU
-
-   .. code-block:: console
-
-      (locpix-env) $ pip install pytorch
-      (locpix-env) $ git clone https://github.com/oubino/cellpose
-      (locpix-env) $ cd cellpose
-      (locpix-env) $ pip install .
-
-
-Perform Cellpose segmentation on our localisation dataset.
-
-To run the script using the GUI, run
+If you have a GPU
 
 .. code-block:: console
 
-   (locpix-env) $ cellpose_eval
+   (locpix-env) $ pip install torch torchvision --extra-index-url https://download.pytorch.org/whl/cu117
+   (locpix-env) $ git clone https://github.com/oubino/cellpose
+   (locpix-env) $ cd cellpose
+   (locpix-env) $ pip install .
 
-To run the script without a GUI -i and -c flags should be specified
+If you don't have a GPU
 
 .. code-block:: console
 
-   (locpix-env) $ cellpose_eval -i path/to/project/directory -c path/to/config/file
+   (locpix-env) $ pip install pytorch
+   (locpix-env) $ git clone https://github.com/oubino/cellpose
+   (locpix-env) $ cd cellpose
+   (locpix-env) $ pip install .
 
 
-Cellpose segmentation (training)
-^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+To evaluate Cellpose model on the localisation dataset without any retraining on your dataset run the script with -i and -c flags specified
 
-   Need to activate extra requirements - these are big and not included in initial install.
+   .. code-block:: console
 
-   Note that if you have a GPU this will speed this up.
+      (locpix-env) $ cellpose_eval -i path/to/project/directory -c path/to/config/file
 
-   Note we modified Cellpose to fit in with our analysis, therefore you need to install our forked repository - note below will clone the Cellpose repository to wherever you are located
+To retrain first then evaluate we instead
 
-   If you have a GPU
+   Prepare data for training
 
    .. code-block:: console
 
-      (locpix-env) $ pip install torch torchvision --extra-index-url https://download.pytorch.org/whl/cu117
-      (locpix-env) $ git clone https://github.com/oubino/cellpose
-      (locpix-env) $ cd cellpose
-      (locpix-env) $ pip install .
+      (locpix-env) $ train_prep -i path/to/project/directory -c path/to/config/file
 
-   If you don't have a GPU
+   Train cellpose
 
    .. code-block:: console
 
-      (locpix-env) $ pip install pytorch
-      (locpix-env) $ git clone https://github.com/oubino/cellpose
-      (locpix-env) $ cd cellpose
-      (locpix-env) $ pip install .
-
-
-Prepare data for training
-
-.. code-block:: console
-
-   (locpix-env) $ cellpose_train_prep -i path/to/project/directory -c path/to/config/file
-
-Train cellpose (using their scripts)
-
-.. code-block:: console
-
-   (locpix-env) $ python -m cellpose --train --dir path/to/project/directory/cellpose_train/train --test_dir path/to/project/directory/cellpose_train/test --pretrained_model LC1 --chan 0 --chan2 0 --learning_rate 0.1 --weight_decay 0.0001 --n_epochs 10 --min_train_masks 1 --verbose
-
-Evaluate cellpose
-
-.. code-block:: console
-
-   (locpix-env) $ cellpose_eval -i path/to/project/directory -c path/to/config/file -u -o cellpose_train_eval
+      (locpix-env) $ cellpose_train -i path/to/project/directory -ct path/to/config/train_file -ce path/to/config/eval_file
 
 
 Ilastik segmentation

diff --git a/docs/source/index.rst b/docs/source/index.rst
@@ -10,8 +10,7 @@ This includes the following functionality:
 #. :ref:`get-markers` : Labelling histogram with seeds for watershed algorithm
 #. Segmentation:
    #. :ref:`classic-segmentation` : Use classic method to segment histograms to extract relevant localisations
-   #. :ref:`cellpose-segmentation-train` : Use Cellpose method to train cellpose to extract relevant localisations
-   #. :ref:`cellpose-segmentation-eval` : Use Cellpose method to segment histograms to extract relevant localisations
+   #. :ref:`cellpose-segmentation` : Use Cellpose method to segment histograms to extract relevant localisations
    #. :ref:`ilastik-segmentation` : Use Ilastik method to segment histograms to extract relevant localisations
 #. :ref:`membrane-performance` : Performance metrics calculation based on the localisations (not the histograms!)
 

diff --git a/docs/source/user_guide/quickstart.rst b/docs/source/user_guide/quickstart.rst
@@ -15,12 +15,10 @@ will not be accessed again!
 Usage configuration
 -------------------
 
-Each script can be run with a GUI, but can also be run in headless mode.
-
-In headless mode each script needs a configuration file (.yaml file), which should be
+Each script needs a configuration file (.yaml file), which should be
 specified using the -c flag.
 
-Each configuration used, whether run in GUI or headless mode will be saved in the project directory.
+Each configuration used will be saved in the project directory.
 
 The templates for the configuration files can be found in :ref:`templates`
 
@@ -58,13 +56,7 @@ This script preprocesses the input .csv data for later use AND **must be run fir
 This script will take in .csv files, and convert them to .parquet files,
 while also wrangling the data into our data format.
 
-To run the script using the GUI, run
-
-.. code-block:: console
-
-   (locpix-env) $ preprocess
-
-To run the script without a GUI -i -c and -o flags should be specified
+To run the script -i -c and -o flags should be specified
 
 .. code-block:: console
 
@@ -78,13 +70,7 @@ Annotate
 
 This script allows for manual segmentation of the localisations.
 
-To run the script using the GUI, run
-
-.. code-block:: console
-
-   (locpix-env) $ annotate
-
-To run the script without a GUI -i and -c flags should be specified
+To run the script -i and -c flags should be specified
 
 .. code-block:: console
 
@@ -101,13 +87,7 @@ Get markers
 
 This script allows for labelling the localisation image with a marker to represent the cells.
 
-To run the script using the GUI, run
-
-.. code-block:: console
-
-   (locpix-env) $ get_markers
-
-To run the script without a GUI -i and -c flags should be specified
+To run the script -i and -c flags should be specified
 
 .. code-block:: console
 
@@ -121,13 +101,7 @@ Classic segmentation
 
 Perform classic segmentation on our localisation dataset.
 
-To run the script using the GUI, run
-
-.. code-block:: console
-
-   (locpix-env) $ classic
-
-To run the script without a GUI -i and -c flags should be specified
+To run the script without -i and -c flags should be specified
 
 .. code-block:: console
 
@@ -166,22 +140,31 @@ Cellpose segmentation
         (locpix-env) $ cd cellpose
         (locpix-env) $ pip install .
 
-Perform Cellpose segmentation on our localisation dataset.
+Perform Cellpose segmentation on our without any retraining on your dataset run the script with -i and -c flags specified
 
-To run the script using the GUI, run
+   .. code-block:: console
 
-.. code-block:: console
+      (locpix-env) $ cellpose_eval -i path/to/project/directory -c path/to/config/file
 
-   (locpix-env) $ cellpose
+To retrain first then evaluate we instead
 
-To run the script without a GUI -i and -c flags should be specified
+   Prepare data for training
 
-.. code-block:: console
+   .. code-block:: console
+
+      (locpix-env) $ train_prep -i path/to/project/directory -c path/to/config/file
+
+   Train cellpose
+
+   .. code-block:: console
+
+      (locpix-env) $ cellpose_train -i path/to/project/directory -ct path/to/config/train_file -ce path/to/config/eval_file
 
-   (locpix-env) $ cellpose -i path/to/project/directory -c path/to/config/file
 
 **API**
-:py:mod:`locpix.scripts.img_seg.cellpose`
+:py:mod:`locpix.scripts.img_seg.train_prep`
+:py:mod:`locpix.scripts.img_seg.cellpose_eval`
+:py:mod:`locpix.scripts.img_seg.cellpose_train`
 
 Ilastik segmentation
 ^^^^^^^^^^^^^^^^^^^^