re formatting the api to contain the image bytes in json

app.py

@@ -4,38 +4,14 @@ from flask_cors import CORS
 from dataset.iris import iris
 from opts import options
 
-import os
-
 # using the iris data set for every algorithm
+# just for simplicity sake
 X, y = iris()
 
-app = Flask(
-    __name__,
-    template_folder="templates",
-)
+app = Flask(__name__)
 
 CORS(app, origins="*")
 
-UPLOAD_FOLDER = os.getcwd() + "/plots"
-
-
-@app.route("/", methods=["GET"])
-def index():
-    return render_template("index.html")
-
-
-@app.route("/plots/<plt_key>", methods=["GET"])
-def get_plot(plt_key):
-    filename = f"{plt_key}.png"
-    filepath = os.path.join(UPLOAD_FOLDER, filename)
-
-    if os.path.isfile(filepath):
-        with open(filepath, "rb") as file:
-            plot_bytes = file.read()
-        return plot_bytes, 200, {"Content-Type": "image/png"}
-    else:
-        return "Plot not found", 404
-
 
 @app.route("/neural-network", methods=["POST"])
 def neural_network():

cluster/clusterer.py

@@ -5,7 +5,7 @@ from typing import Callable
 @dataclass
 class Clusterer:
     cluster_func: Callable
-    plot_key = None
+    plt_data = None
 
     def eval(
         self,
@@ -14,6 +14,9 @@ class Clusterer:
     ) -> None:
         ...
 
+    def set_plot_data(self, plt_data):
+        self.plt_data = plt_data
+
     @classmethod
     def from_dict(cls, dct: dict):
         return cls(**dct)

cluster/kmeans.py

@@ -74,5 +74,5 @@ class Kmeans(Clusterer):
             "k": self.k,
             "max_iter": self.max_iter,
             "clusters": cluster_data,
-            "plot_key": self.plot_key,
+            "plt_data": self.plt_data,
         }

cluster/main.py

@@ -17,5 +17,6 @@ def main(
     alg = cluster_alg.from_dict(args)
 
     alg.build(X)
-    plot(clusterer=alg, X=X)
+    plt_data = plot(clusterer=alg, X=X)
+    alg.set_plot_data(plt_data)
     return alg.to_dict(X)

cluster/plot.py

@@ -1,18 +1,14 @@
 import matplotlib
 import matplotlib.pyplot as plt
 import seaborn as sns
-from plt_id import generate_image_key
-import os
+import io
 
 
 matplotlib.use("Agg")
 sns.set()
 
-UPLOAD_FOLDER = os.getcwd() + "/plots"
-os.makedirs(UPLOAD_FOLDER, exist_ok=True)
 
-
-def plot(clusterer, X) -> None:
+def plot(clusterer, X):
     cluster_data = clusterer.to_dict(X)["clusters"]
     fig, ax = plt.subplots(figsize=(8, 6))
     for cluster in cluster_data:
@@ -35,10 +31,9 @@ def plot(clusterer, X) -> None:
     ax.set_ylabel("Normalized Petal Length (cm)")
     ax.set_xlabel("Normalized Petal Length (cm)")
 
-    image_key = generate_image_key()
-
-    plot_filename = os.path.join(UPLOAD_FOLDER, f"{image_key}.png")
-    fig.savefig(plot_filename, format="png")
-    plt.close(fig)
+    # Save the plot to a BytesIO buffer
+    buffer = io.BytesIO()
+    plt.savefig(buffer, format='png')
+    buffer.seek(0)
 
-    clusterer.plot_key = image_key
+    return buffer.read()

example/kmeans.py

@@ -8,8 +8,8 @@ import seaborn as sns
 ENDPOINT: str = "http://127.0.0.1:5000/"
 
 request_params = {
-    "algorithm": "kmeans-clustering",
     "arguments": {
+        "clusterer": "kmeans-clustering",
         "k": 3,
         "max_iter": 100,
     },
@@ -27,31 +27,5 @@ r = requests.post(
 )
 
 
-def plot():
-    cluster_data = r.json()["clusters"]
-    # plot the clusters and data points
-    fig, ax = plt.subplots(figsize=(8, 6))
-
-    sns.set()
-    for cluster in cluster_data:
-        sns.scatterplot(
-            x=[point[0] for point in cluster["points"]],
-            y=[point[1] for point in cluster["points"]],
-            label=f"Cluster {cluster['cluster_id']}",
-            ax=ax,
-        )
-        ax.scatter(
-            x=cluster["centroid"][0],
-            y=cluster["centroid"][1],
-            marker="x",
-            s=100,
-            linewidth=2,
-            color="black"
-        )
-    ax.set_title("K-means Clustering")
-    ax.legend()
-    plt.show()
-
-
 if __name__ == "__main__":
-    plot()
+    print(r.json())

example/neural_network.py

@@ -1,5 +1,3 @@
-import matplotlib.pyplot as plt
-import seaborn as sns
 import requests
 import json
 
@@ -25,17 +23,5 @@ r = requests.post(
     data=json.dumps(request_params),
 )
 
-model = r.json()
-
-
-def plot():
-    sns.set()
-    plt.plot(model["loss_history"])
-    plt.xlabel("Epoch")
-    plt.ylabel("Loss")
-    plt.title("Loss History")
-    plt.show()
-
-
 if __name__ == "__main__":
-    plot()
+    print(r.json()["plt_data"])

neural_network/main.py

@@ -3,7 +3,6 @@ import numpy as np
 
 from neural_network.opts import activation
 from neural_network.backprop import bp
-from neural_network.plot import plot
 
 
 def init(X: np.array, hidden_size: int):
@@ -19,7 +18,7 @@ def init(X: np.array, hidden_size: int):
     }
 
 
-def main( 
+def main(
     X: np.array,
     y: np.array,
     args,
@@ -48,6 +47,5 @@ def main(
         y_test=y_test,
     )
 
-    plot(model=model)
-
+    model.plot()
     return model.to_dict()

neural_network/neural_network.py

@@ -1,6 +1,15 @@
 from dataclasses import dataclass, field
+from matplotlib import pyplot as plt
+import matplotlib
+import seaborn as sns
 from typing import Callable
 import numpy as np
+import base64
+import binascii
+import io
+
+sns.set()
+matplotlib.use("Agg")
 
 
 @dataclass
@@ -19,7 +28,7 @@ class NeuralNetwork:
     loss_history: list = field(
         default_factory=lambda: [],
     )
-    plot_key = None
+    plt_data = None
 
     def predict(self, x: np.array) -> np.array:
         n1 = self.compute_node(x, self.w1, self.b1, self.activation_func)
@@ -31,6 +40,29 @@ class NeuralNetwork:
     def eval(self, X_test, y_test) -> None:
         self.mse = np.mean((self.predict(X_test) - y_test) ** 2)
 
+    def set_plot_data(self, plot_data):
+        self.plt_data = plot_data
+
+    def plot(self):
+        sns.set()
+        fig, ax = plt.subplots()
+        sns.lineplot(
+            x=np.arange(len(self.loss_history)),
+            y=self.loss_history,
+            ax=ax,
+        )
+        plt.ylabel("Loss")
+        plt.xlabel("Epoch")
+        plt.title("Loss / Epoch")
+
+        buffer = io.BytesIO()
+        plt.savefig(buffer, format='png')
+        buffer.seek(0)
+        plt_data = buffer.read()
+        plt_data_encoded = binascii.hexlify(plt_data).decode('utf-8')
+        self.set_plot_data(plt_data_encoded)
+        plt.close()
+
     @staticmethod
     def compute_node(arr, w, b, func) -> np.array:
         return func(np.dot(arr, w) + b)
@@ -41,18 +73,15 @@ class NeuralNetwork:
 
     def to_dict(self) -> dict:
         return {
-            "w1": self.w1.tolist(),
-            "w2": self.w2.tolist(),
-            "b1": self.b1.tolist(),
-            "b2": self.b2.tolist(),
+            # "w1": self.w1.tolist(),
+            # "w2": self.w2.tolist(),
+            # "b1": self.b1.tolist(),
+            # "b2": self.b2.tolist(),
             "epochs": self.epochs,
             "learning_rate": self.learning_rate,
             "activation_func": self.activation_func.__name__,
             "func_prime": self.func_prime.__name__,
             "hidden_size": self.hidden_size,
             "mse": self.mse,
-            # not returning this because we are making our own
-            # plots and this can be a lot of data
-            # "loss_history": self.loss_history,
-            "plot_key": self.plot_key,
+            "plt_data": self.plt_data,
         }

neural_network/plot.py

@@ -1,32 +0,0 @@
-import numpy as np
-import seaborn as sns
-import matplotlib
-import matplotlib.pyplot as plt
-from plt_id import generate_image_key
-import os
-
-matplotlib.use("Agg")
-
-UPLOAD_FOLDER = os.getcwd() + "/plots"
-os.makedirs(UPLOAD_FOLDER, exist_ok=True)
-
-
-def plot(model) -> None:
-    sns.set()
-    fig, ax = plt.subplots()
-    sns.lineplot(
-        x=np.arange(len(model.loss_history)),
-        y=model.loss_history,
-        ax=ax,
-    )
-    plt.ylabel("Loss")
-    plt.xlabel("Epoch")
-    plt.title("Loss / Epoch")
-
-    image_key = generate_image_key()
-
-    plot_filename = os.path.join(UPLOAD_FOLDER, f"{image_key}.png")
-    fig.savefig(plot_filename, format="png")
-    plt.close(fig)
-
-    model.plot_key = image_key

plt_id.py

@@ -1,5 +0,0 @@
-import uuid
-
-
-def generate_image_key() -> str:
-    return str(uuid.uuid4())

templates/index.html

@@ -1,12 +0,0 @@
-<!DOCTYPE html>
-<html>
-  <head>
-      <title>Data Mining From Scratch Backend</title>
-  </head>
-  <body>
-      <h1>Data Mining From Scratch Backend</h1>
-        <p>This is the backend source code for Data Mining From Scratch.</p>
-          <p>You can find the code on <a href="https://github.com/Jensen-holm/Data-Mining-From-Scratch-Backend">GitHub</a>.</p>
-  </body>
-</html>
-