[init] initial commit

.gitignore

@@ -0,0 +1 @@
+.venv

Dockerfile

@@ -0,0 +1,17 @@
+FROM alpine:latest
+
+MAINTAINER Klesh Wong <klesh@kleshwong.com>
+LABEL description="Simple verification-code solver"
+
+WORKDIR /data
+
+RUN sed -i 's|dl-cdn.alpinelinux.org|mirrors.aliyun.com|g' /etc/apk/repositories
+RUN apk update --no-cache \
+    && apk add --update --no-cache tesseract-ocr py3-pip py3-numpy py3-pillow \
+    && rm -rf /var/cache/apk/*
+
+ADD requirements.txt /data/requirements.txt
+RUN pip3 install -r /data/requirements.txt --index-url=https://mirrors.aliyun.com/pypi/simple
+
+ADD main.py /data/main.py
+CMD python3 /data/main.py

main.py

@@ -0,0 +1,180 @@
+import pytesseract
+import base64
+import numpy as np
+from flask import Flask, request
+from flasgger import Swagger
+from io import BytesIO
+from PIL import Image
+from tempfile import NamedTemporaryFile
+from typing import Tuple
+from os import environ
+
+
+app = Flask(__name__)
+app.config['SWAGGER'] = {
+    'title': 'Simple verification-code solver',
+}
+
+Swagger(app)
+
+
+def remove_noise(img_array: np.ndarray, dominant_color: Tuple[int, int, int], threshold=10):
+    img_array = img_array.copy()
+    h, w, c = img_array.shape
+    visited = set()
+
+    grid = [
+        (-1, -1),
+        (-1, 0),
+        (-1, +1),
+        (0, -1),
+        (0, +1),
+        (+1, -1),
+        (+1, 0),
+        (+1, +1)
+    ]
+
+    def find_connected(y: int, x: int):
+        if y < 0 or y >= h or x < 0 or x >= w:
+            return
+        if (y, x) in visited:
+            return
+        visited.add((y, x))
+        if np.array_equal(img_array[y, x], dominant_color):
+            return
+
+        connected = [(y, x)]
+        for dy, dx in grid:
+            next_connected = find_connected(dy+y, dx+x)
+            if next_connected:
+                connected += next_connected
+        return connected
+
+    for y in range(h):
+        for x in range(w):
+            pixel = img_array[y, x]
+            if pixel[0] != pixel[1] or pixel[1] != pixel[2]:
+                img_array[y, x] = dominant_color
+
+    for y in range(h):
+        for x in range(w):
+            connected = find_connected(y, x)
+            if not connected or len(connected) > threshold:
+                continue
+            for pos in connected:
+                img_array[pos[0], pos[1]] = dominant_color
+
+    return img_array
+
+
+def sanitize(pil_img: Image) -> np.ndarray:
+    data = np.asarray(pil_img)
+    clip = data[1:-1, 1:-1]
+    count_colors = sorted(pil_img.getcolors(), key=lambda cc: cc[0])
+    dominant_color = np.asarray(count_colors[-1][1])
+    return remove_noise(clip, dominant_color)
+
+
+@app.route('/verification-code', methods=['POST'])
+def solve_verification_code():
+    """
+    Solve verification code
+    ---
+    tags:
+      - Verification Code
+    parameters:
+      - in: formData
+        name: image
+        type: file
+        required: true
+        description: Image file
+      - in: formData
+        name: timeout
+        type: int
+        required: false
+        description: Timeout
+      - in: header
+        name: X-Key
+        type: string
+        required: true
+    responses:
+      500:
+        description: Error message
+      200:
+        description: text
+    """
+
+    x_key = request.headers.get('X-Key')
+    if x_key != environ.get('X_KEY'):
+        return "Unauthorized", 401
+
+    if 'image' not in request.files:
+        return "No file uploaded", 400
+
+    image_file = request.files['image']
+
+    if image_file.filename == '':
+        return "No file name", 400
+
+    timeout = 300
+    if "timeout" in request.form and int(request.form['timeout']) > 0:
+        timeout = int(request.form['timeout'])
+
+    img = Image.open(BytesIO(image_file.read()))
+    ary = sanitize(img)
+    txt = pytesseract.image_to_string(ary, timeout=timeout)
+
+    return txt
+
+
+@app.route('/verification-code/base64', methods=['POST'])
+def sove_verification_code_base64():
+    """
+    Solve verification code from base64-encoded image
+    ---
+    tags:
+      - Verification Code
+    parameters:
+      - in: formData
+        name: image
+        type: string
+        required: true
+        description: Base64-encoded image
+      - in: formData
+        name: timeout
+        type: int
+        required: false
+        description: Timeout
+      - in: header
+        name: X-Key
+        type: string
+        required: true
+    responses:
+      500:
+        description: Error message
+      200:
+        description: text
+    """
+
+    x_key = request.headers.get('X-Key')
+    if x_key != environ.get('X_KEY'):
+        return "Unauthorized", 401
+
+    image = request.form.get('image')
+    if not image:
+        return 'No image data', 400
+
+    timeout = 300
+    if "timeout" in request.form and int(request.form['timeout']) > 0:
+        timeout = int(request.form['timeout'])
+
+    img = Image.open(BytesIO(base64.b64decode(image)))
+    ary = sanitize(img)
+    txt = pytesseract.image_to_string(ary, timeout=timeout)
+
+    return txt
+
+
+
+if __name__ == "__main__":
+    app.run(debug=environ.get("FLASK_ENV") == 'development', host='0.0.0.0', port=8000)

requirements.txt

@@ -0,0 +1,6 @@
+pillow
+pytesseract
+flask
+flasgger
+numpy
+waitress