From 2d55c3a5f5d17b7c5e9d86ac5f7b7a96c84f7eae Mon Sep 17 00:00:00 2001
From: merschie <f-mersch@web.de>
Date: Thu, 26 Oct 2023 13:54:00 +0200
Subject: [PATCH] get Namelist from client
---
src/preprocessor.py | 2 +-
src/webserver.py | 152 ++++++++++++--------------------------------
2 files changed, 42 insertions(+), 112 deletions(-)
diff --git a/src/preprocessor.py b/src/preprocessor.py
index ea634b8..7b21d81 100644
--- a/src/preprocessor.py
+++ b/src/preprocessor.py
@@ -165,7 +165,7 @@ class Preprocessor:
#cut out first horizontal line if it is white
img = 255-img
try:
- while np.sum(img[0,:]) < 2048:
+ while np.sum(img[0,:]) < 1024:
img = img[1:,:]
except:
return np.ones((ht, ht))
diff --git a/src/webserver.py b/src/webserver.py
index 7bd016c..e1c19e4 100644
--- a/src/webserver.py
+++ b/src/webserver.py
@@ -1,67 +1,66 @@
from flask import Flask, request, jsonify
-
import main as htr
import model as htr_model
import dataloader_iam as htr_data_loader
import preprocessor as htr_preprocessor
import numpy as np
import base64
-import time
+import json
app = Flask(__name__)
-image_size = 32
-
model_name = htr_model.Model(htr.char_list_from_file(), htr_model.DecoderType.LexiconSearch, must_restore=True)
model_name.setup_ctc
-csv_path = '../tns.csv'
char_list = htr.char_list_from_file()
chars = ''.join(char_list)
-#word_chars = open('../model/wordCharList.txt').read().splitlines()[0]
-matrikel_numbers = []
-
-@app.route('/getNames', methods=['GET'])
-def getNames():
- return jsonify(matrikel_numbers)
-
+def image_to_base64(processed_image):
+ processed_image = processed_image + 0.5
+ processed_image = processed_image * 255
+ #rotate image 90 degrees
+ processed_image = np.rot90(processed_image,3)
+ #mirror image
+ image = np.fliplr(processed_image)
+ height, width = image.shape
+ cut = np.reshape(image,(height*width))
+ cut = np.append(cut,height)
+ cut = np.append(cut,width)
+ cut = cut.astype(np.uint64)
+ array_bytes = cut.tobytes()
+ image_base64 = base64.b64encode(array_bytes).decode('utf-8')
+ return image_base64
-@app.route('/predictNachname', methods=['POST'])
-def predictNach():
- image_array = np.frombuffer(request.data, dtype=np.uint64)
+def base64_to_image(processed_image):
+ image=base64.b64decode(processed_image.encode('utf-8'))
+ image_array = np.frombuffer(image, dtype=np.uint64)
h=image_array[-2]
w=image_array[-1]
image_array = image_array[:-2]
image_array = image_array.reshape((h, w))
+ return image_array
+
+
+
+@app.route('/predictNachname', methods=['POST'])
+def predictNach():
+ image_data=json.loads(request.data)['image']
+
+ image_array=base64_to_image(image_data)
+
preprocessor = htr_preprocessor.Preprocessor(htr.get_img_size(), dynamic_width=True, padding=16)
processed_image = preprocessor.process_img(image_array)
batch = htr_data_loader.Batch([processed_image], None, 1)
#change corpus for name
- model_name.corpus = open('../data/Nachname.txt').read().split()[:200]
+ names=json.loads(request.data)['names']
+ print(names)
+ model_name.corpus = names.split(' ')
- #check time of infer_batch
recognized, probability = htr_model.Model.infer_batch(model_name, batch)
-
- processed_image = processed_image + 0.5
- processed_image = processed_image * 255
- #rotate image 90 degrees
- processed_image = np.rot90(processed_image,3)
- #mirror image
- processed_image = np.fliplr(processed_image)
- height, width = processed_image.shape
- image = np.reshape(processed_image,(height*width))
- image = np.append(image,height)
- image = np.append(image,width)
- image = image.astype(np.uint64)
- array_bytes = image.tobytes()
- image_base64 = base64.b64encode(array_bytes).decode('utf-8')
-
-
-
+ image_base64 = image_to_base64(processed_image)
result = {
'recognized': recognized,
@@ -73,36 +72,21 @@ def predictNach():
@app.route('/predictVorname', methods=['POST'])
def predictVor():
- image_array = np.frombuffer(request.data, dtype=np.uint64)
- h=image_array[-2]
- w=image_array[-1]
- image_array = image_array[:-2]
- image_array = image_array.reshape((h, w))
+ image_data=json.loads(request.data)['image']
+
+ image_array=base64_to_image(image_data)
+
preprocessor = htr_preprocessor.Preprocessor(htr.get_img_size(), dynamic_width=True, padding=16)
processed_image = preprocessor.process_img(image_array)
batch = htr_data_loader.Batch([processed_image], None, 1)
-
#change corpus for name
- model_name.corpus = open('../data/Vorname.txt').read().split()[:200]
- recognized, probability = htr_model.Model.infer_batch(model_name, batch)
+ names=json.loads(request.data)['names']
+ model_name.corpus = names.split(' ')
+ recognized, probability = htr_model.Model.infer_batch(model_name, batch)
-
- processed_image = processed_image + 0.5
- processed_image = processed_image * 255
- #rotate image -90 degrees
- processed_image = np.rot90(processed_image,3)
- #mirror image
- processed_image = np.fliplr(processed_image)
- height, width = processed_image.shape
- image = np.reshape(processed_image,(height*width))
- image = np.append(image,height)
- image = np.append(image,width)
- image = image.astype(np.uint64)
- array_bytes = image.tobytes()
- image_base64 = base64.b64encode(array_bytes).decode('utf-8')
-
+ image_base64 = image_to_base64(processed_image)
result = {
'recognized': recognized,
@@ -110,60 +94,6 @@ def predictVor():
}
return jsonify(result)
-def replace_umlauts(text):
- text = text.replace('ä', 'a')
- text = text.replace('ö', 'o')
- text = text.replace('ü', 'u')
- text = text.replace('Ä', 'A')
- text = text.replace('Ö', 'O')
- text = text.replace('Ü', 'U')
- text = text.replace('ß', 's')
- text = text.replace('é', 'e')
- text = text.replace('è', 'e')
- text = text.replace('ê', 'e')
- text = text.replace('à', 'a')
- text = text.replace('â', 'a')
- text = text.replace('á', 'a')
- text = text.replace('ô', 'o')
- text = text.replace('û', 'u')
- text = text.replace('ç', 'c')
- text = text.replace('î', 'i')
- text = text.replace('ï', 'i')
- text = text.replace('ë', 'e')
- text = text.replace('ù', 'u')
- text = text.replace(' ', '-')
- return text
-
-
-def split_Student_Names():
- #csv looks like: Vorname;Nachname;Matrikelnummer
- #need to put Vorname in one list, Nachname in another
-
- #create /data/Vorname.txt and /data/Nachname.txt
- vorname_file = open('../data/Vorname.txt', 'w')
- nachname_file = open('../data/Nachname.txt', 'w')
- matrikelnummer_file = open('../data/Matrikelnummer.txt', 'w')
- teilnehmer=[]
- with open(csv_path, 'r') as csv_file:
- lines = csv_file.readlines()
- for line in lines[1:]:
- line = line.split(',')
- vorname = line[2][:-1]
- vorname = replace_umlauts(vorname)
- nachname = line[1]
- nachname = replace_umlauts(nachname)
- matrikelnummer = line[0]
- print(vorname, nachname, matrikelnummer)
- vorname_file.write(vorname + " ")
- nachname_file.write(nachname + " ")
- matrikelnummer_file.write(matrikelnummer + " ")
- teilnehmer.append([vorname,nachname,matrikelnummer])
- return teilnehmer
-
-
if __name__ == '__main__':
- #split csv file into Vorname, Nachname and Matrikelnummer
- matrikel_numbers = split_Student_Names()
-
app.run(debug=False,port=8000)
--
GitLab