feat: neural network interface

src/index.html

@@ -14,6 +14,8 @@
   <!-- Styles -->
   <link rel="stylesheet" href="styles.css">
 
+  <script src="view/ordersView.js"></script>
+
   <!-- logic -->
   <script src="configuration.js"></script>
   <script src="utilities.js"></script>
@@ -27,7 +29,6 @@
 
   <script src="view/floor.js"></script>
   <script src="view/grid.js"></script>
-  <script src="view/ordersView.js"></script>
   <script src="logic/pathfinding.js"></script>
   <script src="view/timeController.js"></script>
   <script src="view/agentController.js"></script>
@@ -316,11 +317,10 @@
         </div>
       </div>
       <div class="line gaps">
-        <div class="response">
-          <!-- PYTHON RESPONSE HERE -->
-        </div>
+        <div class="response"></div>
         <button class="clear" onClick="OrdersView.instance.clear()">CLEAR</button>
         <button class="order" onClick="OrdersView.instance.order()">ORDER</button>
+        <button class="order" onClick="OrdersView.instance.addOrderByLastRecognizedString()">OK</button>
       </div>
     </div>
   </div>

src/logic/agent.js

@@ -10,6 +10,7 @@ class Agent extends AgentController {
     const cycle = async () => {
       try {
         const jobRequest = Products.instance.getJobRequest();
+        console.log(jobRequest);
         if (jobRequest) {
           await this.takeFromStore(jobRequest.product, jobRequest.amount);
           await this.deliver(jobRequest.x, jobRequest.y);

src/main.py

@@ -1,5 +1,6 @@
 from flask import Flask, redirect, jsonify, request
 import decision_tree
+from neural_network import *
 
 app = Flask(
   __name__,
@@ -31,6 +32,28 @@ def api_predict_shelf():
 
   return jsonify([ n2t['polka'][clf.predict([X]).tolist()[0]] ])
 
+@app.route('/api/neural', methods=['POST'])
+def naural():
+  length = int(request.form['length'])
+  print(length)
+
+  data = list()
+
+  for fileIndex in range(length):
+    file = request.files[f'file-{fileIndex}']
+    file.save(f'user_input/file-{fileIndex}.png')
+    data.append(f'user_input/file-{fileIndex}.png')
+
+  return jsonify(testCase(data))
+
+  # try:
+  #   return("Image uploaded")
+  #   print("Image uploaded")
+  # except Exception as err:
+  #   print("Error occurred")
+  #   print(err)
+  #   return("Error, image not received.")
+
 @app.route('/')
 def index():
   return redirect('/index.html')

src/neural_network.py

@@ -30,8 +30,8 @@ class NeuralNetwork:
         self.weights = (np.random.rand(self.hnodes, self.inodes) - 0.5)
         self.hidden = (np.random.rand(self.onodes, self.hnodes) - 0.5)
         """ używane przy pobieraniu danych o nauczonej sieci, z pliku """
-        # self.weights = np.load(fileWeight)
-        # self.hidden = np.load(fileHidden)
+        self.weights = np.load(fileWeight)
+        self.hidden = np.load(fileHidden)
 
         self.lr = learningGrade
 
@@ -136,23 +136,26 @@ li = []
 ourOwnDataset = []
 record_cache = None
 def testCase(inputWord):
-    len = len(inputWord)
+    length = len(inputWord)
     
     word = ""
-    for i in range(0,len-2):
-        imgArray = imageio.imread(imageFileName, as_gray=True)
+    for i in range(0, length):
+        imgArray = imageio.imread(inputWord[i], as_gray=True)
         imgData = 255 - imgArray.reshape(784)
         imgData = (imgData/255 * 0.99) + 0.01
         #inputWord[i]
-        word = word + recognizeLet(letterNetwork ,imgData)
-    word = word + recognizeNum(digitNetwork, inputWord[-2])
-    word = word + recognizeNum(digitNetwork ,inputWord[-1])
+
+        if i < length - 2:
+            word = word + recognizeLet(letterNetwork, imgData)
+        else:
+            word = word + recognizeNum(digitNetwork, imgData)
 
 
     #assert record_cache.shape == ourOwnDataset[0].shape
     #labelInput = np.asfarray(li)
     #print(labelInput)
     print('slowo: ', word)
+    return word
     pass
 
 
@@ -234,5 +237,5 @@ print('slowo: ', word)
 
 
 #####################################      URUCHOMIENIE TRENINGU
-trainNetwork(letterNetwork, "Lweights_test.npy", "Lhidden_test.npy", let_train_images, let_train_labels)
+# trainNetwork(letterNetwork, "Lweights_test.npy", "Lhidden_test.npy", let_train_images, let_train_labels)
 # trainNetwork(digitNetwork, "Dweights_test.npy", "Dhidden_test.npy", let_train_images, let_train_labels)

src/view/ordersView.js

@@ -3,6 +3,9 @@ class OrdersView {
   constructor() {
     OrdersView.instance = this;
 
+    this.orders = [];
+    this.lastRecognizedText = '';
+
     this.addMultipleCanvases(10);
   }
 
@@ -23,6 +26,8 @@ class OrdersView {
               canvas.height = 28;
 
               const ctx = canvas.getContext('2d');
+              ctx.fillStyle = 'white';
+              ctx.fillRect(0, 0, 28, 28);
               ctx.drawImage(img, 0, 0, 28, 28);
 
               canvas.toBlob((blob) => {
@@ -34,13 +39,17 @@ class OrdersView {
           });
         });
       })
-    ).then(async () => {
+    ).then(async (res) => {
+      formData.append('length', res.length);
+
       const response = await fetch('/api/neural', { 
         method: 'POST',
         body: formData
       });
       //const json = await response.json();
-      console.log(await response.text());
+      const res_ = (await response.text()).replace(/[\"]{1,}/g, '');
+      document.querySelector('.response').innerHTML = res_;
+      this.lastRecognizedText = res_;
     });
   }
 
@@ -115,4 +124,29 @@ class OrdersView {
       ctx.stroke();
     }
   }
+
+  addOrderByLastRecognizedString() {
+    this.addOrderByString(this.lastRecognizedText);
+  }
+
+  addOrderByString(str) {
+    const numberStr = str.substring(str.length - 2, str.length);
+    const text = str.substring(0, str.length - 2);
+    const number = parseInt(numberStr);
+
+    this.orders.push({
+      product: Product.REGISTRY[text],
+      amount: number,
+      x: 11,
+      y: 7,
+    });
+
+    this.closeWindow();
+  }
+
+  getOrder() {
+    if (!this.orders.length) return null;
+
+    return this.orders.shift();
+  }
 }