chore: rebase conflict when pushing

6d6f4480 · TheRiPtide · f28b078c · fb8e822e · 6d6f4480 · 6d6f4480
Commit 6d6f4480 authored 3 years ago by TheRiPtide
--- a/notebooks/internal_priming.ipynb
+++ b/notebooks/internal_priming.ipynb
@@ -22,17 +22,22 @@
  },
  {
   "cell_type": "code",
+<<<<<<< HEAD
 <<<<<<< HEAD
   "execution_count": 80,
 =======
   "execution_count": null,
 >>>>>>> d2ef840 (chore: started cnn notebook)
+=======
+   "execution_count": 80,
+>>>>>>> fb8e822ed92fba85e584305fcb18bdf45ad601df
   "outputs": [],
   "source": [
    "# importing the libraries\n",
    "import pandas as pd\n",
    "import numpy as np\n",
 <<<<<<< HEAD
+<<<<<<< HEAD
 <<<<<<< HEAD
    "import matplotlib.pyplot as plt\n",
 =======
@@ -40,6 +45,9 @@
 =======
    "import matplotlib.pyplot as plt\n",
 >>>>>>> 93ea318 (chore: added training function for cnn)
+=======
+    "import matplotlib.pyplot as plt\n",
+>>>>>>> fb8e822ed92fba85e584305fcb18bdf45ad601df
    "\n",
    "# for creating validation set\n",
    "from sklearn.model_selection import train_test_split\n",
@@ -51,6 +59,7 @@
    "# PyTorch libraries and modules\n",
    "import torch\n",
    "from torch.autograd import Variable\n",
+<<<<<<< HEAD
 <<<<<<< HEAD
    "from torch.nn import Linear, ReLU, CrossEntropyLoss, Sequential, MaxPool1d, Module, Softmax, BatchNorm1d, Dropout, Conv1d\n",
    "from torch.optim import Adam\n",
@@ -58,6 +67,10 @@
    "from torch.nn import Linear, ReLU, CrossEntropyLoss, Sequential, Conv2d, MaxPool2d, Module, Softmax, BatchNorm2d, Dropout\n",
    "from torch.optim import Adam, SGD\n",
 >>>>>>> d2ef840 (chore: started cnn notebook)
+=======
+    "from torch.nn import Linear, ReLU, CrossEntropyLoss, Sequential, MaxPool1d, Module, Softmax, BatchNorm1d, Dropout, Conv1d\n",
+    "from torch.optim import Adam\n",
+>>>>>>> fb8e822ed92fba85e584305fcb18bdf45ad601df
    "\n",
    "\n",
    "# adding the nn\n",
@@ -67,6 +80,9 @@
    "\n",
    "        self.cnn_layers = Sequential(\n",
 <<<<<<< HEAD
+<<<<<<< HEAD
+=======
+>>>>>>> fb8e822ed92fba85e584305fcb18bdf45ad601df
    "            # Defining a 1D convolution layer\n",
    "            Conv1d(1, 4, kernel_size=3, stride=1, padding=1),\n",
    "            BatchNorm1d(4),\n",
@@ -81,6 +97,7 @@
    "\n",
    "        self.linear_layers = Sequential(\n",
    "            Linear(4 * 50, 10)\n",
+<<<<<<< HEAD
 =======
    "            # Defining a 2D convolution layer\n",
    "            Conv2d(1, 4, kernel_size=3, stride=1, padding=1),\n",
@@ -97,6 +114,8 @@
    "        self.linear_layers = Sequential(\n",
    "            Linear(4 * 7 * 7, 10)\n",
 >>>>>>> d2ef840 (chore: started cnn notebook)
+=======
+>>>>>>> fb8e822ed92fba85e584305fcb18bdf45ad601df
    "        )\n",
    "\n",
    "    # Defining the forward pass\n",
@@ -106,8 +125,11 @@
    "        x = self.linear_layers(x)\n",
 <<<<<<< HEAD
 <<<<<<< HEAD
+<<<<<<< HEAD
 =======
 >>>>>>> 93ea318 (chore: added training function for cnn)
+=======
+>>>>>>> fb8e822ed92fba85e584305fcb18bdf45ad601df
    "        return x\n",
    "\n",
    "# defining training function\n",
@@ -143,11 +165,14 @@
    "\n",
    "    return loss_train, loss_val"
 <<<<<<< HEAD
+<<<<<<< HEAD
 =======
    "        return x"
 >>>>>>> d2ef840 (chore: started cnn notebook)
 =======
 >>>>>>> 93ea318 (chore: added training function for cnn)
+=======
+>>>>>>> fb8e822ed92fba85e584305fcb18bdf45ad601df
   ],
   "metadata": {
    "collapsed": false,
@@ -171,6 +196,9 @@
  {
   "cell_type": "code",
 <<<<<<< HEAD
+<<<<<<< HEAD
+=======
+>>>>>>> fb8e822ed92fba85e584305fcb18bdf45ad601df
   "execution_count": 81,
   "outputs": [
    {
@@ -250,6 +278,7 @@
    "\n",
    "val_x  = torch.from_numpy(val_x)\n",
    "val_y = torch.from_numpy(val_y)"
+<<<<<<< HEAD
 =======
   "execution_count": null,
   "outputs": [],
@@ -264,6 +293,8 @@
    "\n",
    "# TODO: reshape shape from [n, l] to [n, 1, l]\n"
 >>>>>>> d2ef840 (chore: started cnn notebook)
+=======
+>>>>>>> fb8e822ed92fba85e584305fcb18bdf45ad601df
   ],
   "metadata": {
    "collapsed": false,
@@ -287,6 +318,9 @@
  {
   "cell_type": "code",
 <<<<<<< HEAD
+<<<<<<< HEAD
+=======
+>>>>>>> fb8e822ed92fba85e584305fcb18bdf45ad601df
   "execution_count": 83,
   "outputs": [
    {
@@ -307,6 +341,7 @@
    "# defining the loss function\n",
    "criterion = CrossEntropyLoss()\n",
    "\n",
+<<<<<<< HEAD
 =======
   "execution_count": null,
   "outputs": [],
@@ -324,14 +359,19 @@
 =======
    "\n",
 >>>>>>> 93ea318 (chore: added training function for cnn)
+=======
+>>>>>>> fb8e822ed92fba85e584305fcb18bdf45ad601df
    "# checking if GPU is available\n",
    "if torch.cuda.is_available():\n",
    "    model = model.cuda()\n",
    "    criterion = criterion.cuda()\n",
 <<<<<<< HEAD
 <<<<<<< HEAD
+<<<<<<< HEAD
 =======
 >>>>>>> 93ea318 (chore: added training function for cnn)
+=======
+>>>>>>> fb8e822ed92fba85e584305fcb18bdf45ad601df
    "\n",
    "# defining the number of epochs\n",
    "n_epochs = 25\n",
@@ -344,6 +384,9 @@
    "\n",
    "# training the model\n",
 <<<<<<< HEAD
+<<<<<<< HEAD
+=======
+>>>>>>> fb8e822ed92fba85e584305fcb18bdf45ad601df
    "for epoch in tqdm(range(n_epochs)):\n",
    "    train_loss, val_loss = train()\n",
    "    train_losses.append(train_loss)\n",
@@ -460,6 +503,7 @@
   "outputs": [],
   "source": [
    "torch.save(model.state_dict(), '../models/internal_priming.pth')"
+<<<<<<< HEAD
 =======
    "\n"
 >>>>>>> d2ef840 (chore: started cnn notebook)
@@ -475,6 +519,8 @@
    "plt.legend()\n",
    "plt.show()"
 >>>>>>> 93ea318 (chore: added training function for cnn)
+=======
+>>>>>>> fb8e822ed92fba85e584305fcb18bdf45ad601df
   ],
   "metadata": {
    "collapsed": false,

--- a/src/polyA_classifier.py
+++ b/src/polyA_classifier.py
+<<<<<<< HEAD
 """Module for classifying polyA tails as internal or real."""

+=======
+>>>>>>> fb8e822ed92fba85e584305fcb18bdf45ad601df
 import torch
 from torch.nn import Linear, ReLU, Sequential, MaxPool1d, Module, BatchNorm1d, Conv1d
 import numpy as np
@@ -7,7 +10,10 @@ from typing import Union


 class Net(Module):
+<<<<<<< HEAD
    """Two layer 1D convolutional neural net"""
+=======
+>>>>>>> fb8e822ed92fba85e584305fcb18bdf45ad601df

    def __init__(self):

@@ -30,8 +36,13 @@ class Net(Module):
            Linear(4 * 50, 10)
        )

+<<<<<<< HEAD
    def forward(self, x):
        """Forward pass function."""
+=======
+    # Defining the forward pass
+    def forward(self, x):
+>>>>>>> fb8e822ed92fba85e584305fcb18bdf45ad601df

        x = self.cnn_layers(x)
        x = x.view(x.size(0), -1)
@@ -40,16 +51,23 @@ class Net(Module):


 class PolyAClassifier:
+<<<<<<< HEAD
    """Classifier object using the state-dict of a pretrained pytorch model"""
+=======
+>>>>>>> fb8e822ed92fba85e584305fcb18bdf45ad601df

    enum = {
        'A': 0.0,
        'U': 1 / 3,
+<<<<<<< HEAD
        'T': 1 / 3,
+=======
+>>>>>>> fb8e822ed92fba85e584305fcb18bdf45ad601df
        'G': 2 / 3,
        'C': 1.0
    }

+<<<<<<< HEAD
    def __init__(self, model: Module = Net, state_dict_path: str = './models/internal_priming.pth'):
        """Returns a stateless classifier with the model loaded.

@@ -57,6 +75,9 @@ class PolyAClassifier:
            model: An object subclassing the pytorch Module
            state_dict_path: A path to a saved state-dict of said object at a trained state.
        """
+=======
+    def __init__(self, model=Net, state_dict_path='./models/internal_priming.pth'):
+>>>>>>> fb8e822ed92fba85e584305fcb18bdf45ad601df

        self.model = model()
        self.model.load_state_dict(torch.load(state_dict_path))
@@ -73,7 +94,10 @@ class PolyAClassifier:
        Raises:
            TypeError: If sequence is not str or list(str)
            ValueError: If some or all sequences are not of length 200
+<<<<<<< HEAD
            ValueError: If non-allowed letters in string
+=======
+>>>>>>> fb8e822ed92fba85e584305fcb18bdf45ad601df

        """

@@ -91,6 +115,7 @@ class PolyAClassifier:

        enum_seqs = []

+<<<<<<< HEAD
        try:
            for s in sequences:
                enum_sequence = [self.enum[key.upper()] for key in s]
@@ -112,6 +137,25 @@ class PolyAClassifier:
            raise ValueError('Sequences not of length 200')


+=======
+        for s in sequences:
+            enum_sequence = [self.enum[key] for key in s]
+            enum_seqs.append(enum_sequence)
+
+        # convert to ndarray and reshape for pytorch
+        test = np.array(enum_seqs, dtype=np.float32)
+
+        try:
+            test_shape = test.shape
+            test = test.reshape(test_shape[0], 1, test_shape[1])
+
+            if test_shape[1] != 200:
+                raise ValueError('Sequences not of length 200')
+
+        except IndexError:
+
+            raise ValueError('Not all sequences of length 200')
+>>>>>>> fb8e822ed92fba85e584305fcb18bdf45ad601df

        test = torch.from_numpy(test)

@@ -129,4 +173,29 @@ class PolyAClassifier:

        else:

+<<<<<<< HEAD
            return predictions.tolist()
+=======
+            return predictions
+
+
+if __name__ == '__main__':
+
+    mod = PolyAClassifier(state_dict_path='../models/internal_priming.pth')
+
+    real_str = 'CGCCGGAAGAACGAAUCUCCCACUGCCCGGGCAUCCAAUGGACUUCAUAGGAAUGGCAGCUGAUAACACCGCCCCCUGUGGCGCGCCAGAGGGCGCGCUUCGUGUAGGCUUCGAUGUCGCGGUAAAAUUCUUGGAUUAAAGAAGGGGCCCUGUGGUAGCAAGUUUUUUAUUCUGUGGGCGCUCUUACGCGUGUAUUGUCU'
+    fake_str = 'GUUUGAGGCGCAUGACGCGUUUCGGGGGCCUUGCGUCGCCCACGCCGGCGUUCUCUUUAAAAGGAGCAACGACACCACGCCCCAUGGACCAUGCCGCAGGGUGAACGUCGUCCCGCAACUGCCGUGCACCCGUCAAAAGGAGGCGUCUUCAAAAAAAAAACAAAAUAAAAACACAUACCGCGGCGCGUAUUAGAGCGGCG'
+    list_test = [real_str, fake_str]
+
+    pred = mod.classify(real_str)
+
+    print(pred)
+
+    pred = mod.classify(fake_str)
+
+    print(pred)
+
+    pred = mod.classify(list_test)
+
+    print(pred)
+>>>>>>> fb8e822ed92fba85e584305fcb18bdf45ad601df