feat: added local tensorflow image tagging. You can choose between th…

…at or cloud based clarifai!
elblogbruno · Feb 1, 2021 · 648633e · 648633e
1 parent b6d489e
commit 648633e
Show file tree

Hide file tree

Showing 15 changed files with 252 additions and 149 deletions.
diff --git a/.gitignore b/.gitignore
@@ -1,6 +1,9 @@
 # Created by https://www.toptal.com/developers/gitignore/api/images,video
 # Edit at https://www.toptal.com/developers/gitignore?templates=images,video
 
+/Python Server/app/.idea
+/Python Server/app/image_tagging/temp_image_folder
+/Python Server/app/__pycache__
 *.zip
 *.log
 *.json

diff --git a/Python Server/app/NotionAI.py b/Python Server/app/NotionAI.py
@@ -4,21 +4,24 @@
 import validators
 import os
 
-from custom_errors import OnImageNotFound, OnUrlNotValid, EmbedableContentNotFound, NoTagsFound, OnTokenV2NotValid
+from custom_errors import OnImageNotFound, OnUrlNotValid, EmbedableContentNotFound, NoTagsFound
 
 import requests
 import json
 
-from ClarifaiAI import *
 from uuid import uuid1
 from random import choice
 from time import sleep
 
 import threading
 
+from image_tagging.image_tagging import ImageTagging
+
+
 class NotionAI:
     def __init__(self, logging):
-        print("Init NotionAI")
+        logging.info("Initiating NotionAI Class.")
+
         if os.path.isfile('data.json'):
             print("Initiating with a found config file.")
             logging.info("Initiating with a found config file.")
@@ -28,36 +31,37 @@ def __init__(self, logging):
             print("You should go to the homepage and set the config.")
             logging.info("You should go to the homepage and set the config.")
 
-    def run(self,logging):
-        # Obtain the `token_v2` value by inspecting your browser cookies on a logged-in session on Notion.so
+    def run(self, logging):
         loaded = False
-        options = {}
+        data = {}
         with open('data.json') as json_file:
-            options = json.load(json_file)
+            data = json.load(json_file)
         try:
-            self.client = NotionClient(token_v2=options['token'])
-            mind_page = self.client.get_block(options['url'])
+            self.logging = logging
+
+            self.logging.info("Running notionAI with " + str(data))
+
+            self.data = data
+
+            self.client = NotionClient(token_v2=data['token'])
+
+            mind_page = self.client.get_block(data['url'])
+
             self.mind_id = mind_page.id
 
-            self.options = options
-            self.clarifai = ClarifaiAI(options['clarifai_key'])
+            self.image_tagger = ImageTagging(data, logging)
 
-            cv = self.client.get_collection_view(self.options['url'])
+            cv = self.client.get_collection_view(self.data['url'])
 
             self.collection = cv.collection
-            print("Running notionAI with " + str(self.options))
-
-            self.logging = logging
 
-            self.logging.info("Running notionAI with " + str(self.options))
             loaded = True
         except requests.exceptions.HTTPError:
-            print("Incorrect token V2 from notion")
             self.logging.info("Incorrect token V2 from notion")
         return loaded
 
     def add_url_to_database(self, url, title):
-        print("The url is " + url)
+        self.logging("Adding url {} to mind with title {}".format(url,title))
         self.statusCode = 200  # at start we asume everything will go ok
         try:
             rowId = self.web_clipper_request(url, title)
@@ -68,18 +72,17 @@ def add_url_to_database(self, url, title):
             self.logging.info(invalidUrl)
             self.statusCode = 500
 
-    def add_url_thread(self,rowId):
+    def add_url_thread(self, rowId):
         self.logging.info("Thread %s: starting", rowId)
         row = self.client.get_block(rowId)
         try:
             page_content = self.get_content_from_row(row, 0)
             img_url = self.extract_image_from_content(page_content)
             try:
-                tags = self.clarifai.get_tags(img_url)
+                tags = self.image_tagger.get_tags(img_url)
                 print(tags)
                 self.logging.info(tags)
                 row.AITagsText = tags
-                self.add_new_multi_select_value("AITagsText", tags)
             except NoTagsFound as e:
                 print(e)
                 self.logging.info(e)
@@ -97,49 +100,6 @@ def add_url_thread(self,rowId):
             self.logging.info(e)
         self.logging.info("Thread %s: finishing", rowId)
 
-    def add_new_multi_select_value(self, prop, value, color=None):
-        colors = [
-            "default",
-            "gray",
-            "brown",
-            "orange",
-            "yellow",
-            "green",
-            "blue",
-            "purple",
-            "pink",
-            "red",
-        ]
-
-        """`prop` is the name of the multi select property."""
-        if color is None:
-            color = choice(colors)
-
-        collection_schema = self.collection.get("schema")
-        prop_schema = next(
-            (v for k, v in collection_schema.items() if v["name"] == prop), None
-        )
-        if not prop_schema:
-            raise ValueError(
-                '{} property does not exist on the collection!'.format(prop)
-            )
-        if prop_schema["type"] != "multi_select":
-            raise ValueError('{} is not a multi select property!'.format(prop))
-
-        if "options" not in prop_schema:
-            prop_schema["options"] = []
-
-        dupe = next(
-            (o for o in prop_schema["options"] if o["value"] == value), None
-        )
-        if dupe:
-            raise ValueError(f'"{value}" already exists in the schema!')
-
-        prop_schema["options"].append(
-            {"id": str(uuid1()), "value": value, "color": color}
-        )
-        self.collection.set("schema", collection_schema)
-
     def extract_image_from_content(self, page_content):
         url = " "
         for element in page_content:
@@ -171,7 +131,7 @@ def get_content_from_row(self, row, n):
 
     def web_clipper_request(self, url, title):
         cookies = {
-            'token_v2': self.options['token'],
+            'token_v2': self.data['token'],
         }
 
         headers = {
@@ -200,7 +160,7 @@ def web_clipper_request(self, url, title):
                                      data=data)
             response_text = response.text
             json_response = json.loads(response_text)
-            #self.logging.info(str(json_response))
+            # self.logging.info(str(json_response))
             rowId = json_response['createdBlockIds'][0]
             return rowId
         else:
@@ -229,6 +189,7 @@ def add_image_to_database(self, url, image_src, image_src_url):
         print("The Image is " + image_src + " context: " + image_src_url)
         self.statusCode = 200  # at start we asume everything will go ok
         row = self.collection.add_row()
+
         self.row = row
 
         try:
@@ -238,20 +199,9 @@ def add_image_to_database(self, url, image_src, image_src_url):
             img_block.source = image_src
             row.icon = img_block.source
             row.person = self.client.current_user
-            try:
-                tags = self.clarifai.get_tags(image_src)
-                self.logging.info("tags from image {0} : {1}".format(image_src_url,tags))
-                row.AITagsText = tags
-                self.add_new_multi_select_value("AITags", tags)
-            except NoTagsFound as e:
-                print(e)
-                self.logging.info(e)
-            except ValueError as e:
-                print(e)
-                self.logging.info(e)
-            except Exception as e:
-                print(e)
-                self.logging.info(e)
+            x = threading.Thread(target=self.analyze_image_thread, args=(image_src, row, image_src_url))
+            x.start()
+
         except requests.exceptions.HTTPError as invalidUrl:
             print(invalidUrl)
             self.logging.info(invalidUrl)
@@ -270,21 +220,28 @@ def add_image_to_database_by_post(self, image_src):
             img_block.upload_file(image_src)
             row.icon = img_block.source
             row.person = self.client.current_user
-            try:
-                tags = self.clarifai.get_tags(img_block.source)
-                print(tags)
-                row.AITagsText = tags
-                self.add_new_multi_select_value("AITagsText", tags)
-            except NoTagsFound as e:
-                print(e)
-                self.logging.info(e)
-            except ValueError as e:
-                print(e)
-                self.logging.info(e)
-            except Exception as e:
-                print(e)
-                self.logging.info(e)
+
+            x = threading.Thread(target=self.analyze_image_thread, args=(image_src, row))
+            x.start()
         except requests.exceptions.HTTPError as invalidUrl:
             print(invalidUrl)
             self.logging.info(invalidUrl)
             self.statusCode = 500
+
+    def analyze_image_thread(self, image_src,row, image_src_url="no context"):
+        try:
+            self.logging.info("Image tag Thread %s: starting", row.id)
+            tags = self.image_tagger.get_tags(image_src)
+            self.logging.info("Tags from image {0} : {1}".format(image_src_url, tags))
+            row.AITagsText = tags
+            self.logging.info("Image tag Thread %s: finished", row.id)
+        except NoTagsFound as e:
+            print(e)
+            self.logging.info(e)
+        except ValueError as e:
+            print(e)
+            self.logging.info(e)
+        except Exception as e:
+            print(e)
+            self.logging.info(e)
+
diff --git a/Python Server/app/app.log b/Python Server/app/app.log
@@ -1,27 +1,9 @@
-root - INFO - You should go to the homepage and set the config.
+root - INFO - Initiating NotionAI Class.
+root - INFO - Initiating with a found config file.
+root - INFO - Running notionAI with {'url': 'https://www.notion.so/glassear/d4e43dbfe7244e0a83f18e14506e74ae?v=2b6755006a15418c978d5067180baae0', 'token': 'e6517083645826700a071fdf1644a844b6a9dba7252010c8a99efaae991f990575fd6160ce2c9a30c9ed78057261bb49f3245e227d80c3a8a630f3f4e839e66d90d75c24a26671a29fc072020f6d'}
+root - INFO - Using tensorflow predictor, with delete_after_tagging = True
 root - INFO - Initiating with a found port.json file.
 root - INFO - Using 8080 port
 werkzeug - WARNING -  * Debugger is active!
 werkzeug - INFO -  * Debugger PIN: 207-644-306
 werkzeug - INFO -  * Running on http://0.0.0.0:8080/ (Press CTRL+C to quit)
-root - INFO - Adding url to mind: b'https://github.com/elblogbruno/NotionAI-MyMind' b'elblogbruno/NotionAI-MyMind: This repo uses AI and the wonderful Notion to enable you to add anything on the web to your "Mind" and forget about everything else.'
-werkzeug - INFO - 192.168.1.42 - - [30/Jan/2021 11:09:39] "[35m[1mGET /add_url_to_mind?url=https://github.com/elblogbruno/NotionAI-MyMind&title=elblogbruno/NotionAI-MyMind:%20This%20repo%20uses%20AI%20and%20the%20wonderful%20Notion%20to%20enable%20you%20to%20add%20anything%20on%20the%20web%20to%20your%20%22Mind%22%20and%20forget%20about%20everything%20else. HTTP/1.1[0m" 500 -
-werkzeug - ERROR - 192.168.1.42 - - [30/Jan/2021 11:10:13] code 400, message Bad request version ('��\x13\x01\x13\x02\x13\x03�+�/�,�0̨̩�\x13�\x14\x00\x9c\x00\x9d\x00/\x005\x01\x00\x01\x93��\x00\x00\x00\x17\x00\x00�\x01\x00\x01\x00\x00')
-werkzeug - ERROR - 192.168.1.42 - - [30/Jan/2021 11:10:13] code 400, message Bad HTTP/0.9 request type ('\x16\x03\x01\x02\x00\x01\x00\x01�\x03\x032��\x9e�P�\x92�\x88\x94\x10��\x7f�\x80\x80��,B\x82\x02_��')
-werkzeug - ERROR - 192.168.1.42 - - [30/Jan/2021 11:10:13] code 400, message Bad request version ('Ȧ\x0e��V���\x9c�\x98�x\x83')
-werkzeug - ERROR - 192.168.1.42 - - [30/Jan/2021 11:10:13] code 400, message Bad request version ('2-\x07³�\x0f�w\x13�۬$�\\\x00"JJ\x13\x01\x13\x02\x13\x03�+�/�,�0̨̩�\x13�\x14\x00\x9c\x00\x9d\x00/\x005\x00')
-werkzeug - INFO - 192.168.1.42 - - [30/Jan/2021 11:10:18] "[37mGET / HTTP/1.1[0m" 200 -
-root - INFO - Running notionAI with {'url': 'https://www.notion.so/glassear/d4e43dbfe7244e0a83f18e14506e74ae?v=2b6755006a15418c978d5067180baae0', 'token': 'e6517083645826700a071fdf1644a844b6a9dba7252010c8a99efaae991f990575fd6160ce2c9a30c9ed78057261bb49f3245e227d80c3a8a630f3f4e839e66d90d75c24a26671a29fc072020f6d', 'clarifai_key': '7d13a00d87c5408b96fa94d16ad3c521'}
-werkzeug - INFO - 192.168.1.42 - - [30/Jan/2021 11:10:24] "[37mPOST /handle_data HTTP/1.1[0m" 200 -
-root - INFO - Adding url to mind: b'https://github.com/elblogbruno/NotionAI-MyMind' b'elblogbruno/NotionAI-MyMind: This repo uses AI and the wonderful Notion to enable you to add anything on the web to your "Mind" and forget about everything else.'
-root - INFO - {"type": "block", "blockId": "d4e43dbf-e724-4e0a-83f1-8e14506e74ae", "property": "P#~d", "items": [{"url": "https://github.com/elblogbruno/NotionAI-MyMind", "title": "elblogbruno/NotionAI-MyMind: This repo uses AI and the wonderful Notion to enable you to add anything on the web to your \"Mind\" and forget about everything else."}], "from": "chrome"}
-root - INFO - Thread 058a46df-5f41-401e-be3f-f197d2112859: starting
-werkzeug - INFO - 192.168.1.42 - - [30/Jan/2021 11:10:30] "[37mGET /add_url_to_mind?url=https://github.com/elblogbruno/NotionAI-MyMind&title=elblogbruno/NotionAI-MyMind:%20This%20repo%20uses%20AI%20and%20the%20wonderful%20Notion%20to%20enable%20you%20to%20add%20anything%20on%20the%20web%20to%20your%20%22Mind%22%20and%20forget%20about%20everything%20else. HTTP/1.1[0m" 200 -
-root - INFO - No content available yet 0
-root - INFO - No content available yet 1
-root - INFO - No content available yet 2
-root - INFO - None
-root - INFO - Request failed, status code: 10020
-root - INFO - Thread 058a46df-5f41-401e-be3f-f197d2112859: finishing
-root - INFO - Adding url to mind: b'chrome-extension://ocjdfdihgfjfflddjlkohpeifglmibig/options.html?serverip=http://192.168.1.42:8080/' b'Notion AI My Mind Options!'
-werkzeug - INFO - 192.168.1.42 - - [30/Jan/2021 11:11:59] "[35m[1mGET /add_url_to_mind?url=chrome-extension://ocjdfdihgfjfflddjlkohpeifglmibig/options.html?serverip=http%3A%2F%2F192.168.1.42%3A8080%2F&title=Notion%20AI%20My%20Mind%20Options! HTTP/1.1[0m" 500 -
diff --git a/Python Server/app/image_tagging/__pycache__/image_tagging.cpython-38.pyc b/Python Server/app/image_tagging/__pycache__/image_tagging.cpython-38.pyc
diff --git a/Python Server/app/ClarifaiAI.py → ...ge_tagging/clarifai_tagging/ClarifaiAI.py b/Python Server/app/ClarifaiAI.py → ...ge_tagging/clarifai_tagging/ClarifaiAI.py
diff --git a/Python Server/app/image_tagging/clarifai_tagging/__pycache__/ClarifaiAI.cpython-38.pyc b/Python Server/app/image_tagging/clarifai_tagging/__pycache__/ClarifaiAI.cpython-38.pyc
diff --git a/Python Server/app/image_tagging/image_tagging.py b/Python Server/app/image_tagging/image_tagging.py
@@ -0,0 +1,32 @@
+from image_tagging.clarifai_tagging.ClarifaiAI import ClarifaiAI
+from image_tagging.tensorflow_tagging.tensorflow_tagging import TensorFlowTag
+import json
+
+
+class ImageTagging:
+
+    def __init__(self,data,logging):
+        options = {}
+        with open('options.json') as json_file:
+            options = json.load(json_file)
+        try:
+            self.options = options
+            if options['use_clarifai']:
+                self.predictor = ClarifaiAI(data['clarifai_key'])
+                logging.info("Using clarifai predictor")
+            else:
+                self.predictor = TensorFlowTag(options['delete_after_tagging'])
+                logging.info("Using tensorflow predictor, with delete_after_tagging = {}".format(options['delete_after_tagging']))
+        except FileNotFoundError:
+            logging.info("options.json not found")
+            print("Wrong file or file path")
+
+    def get_tags(self, image_url):
+        self.print_current_detector()
+        return self.predictor.get_tags(image_url)
+
+    def print_current_detector(self):
+        if self.options['use_clarifai']:
+            print("Using clarifai predictor")
+        else:
+            print("Using tensorflow predictor, with delete_after_tagging = {}".format(self.options['delete_after_tagging']))
diff --git a/...Server/app/image_tagging/tensorflow_tagging/__pycache__/tensorflow_tagging.cpython-38.pyc b/...Server/app/image_tagging/tensorflow_tagging/__pycache__/tensorflow_tagging.cpython-38.pyc
diff --git a/Python Server/app/image_tagging/tensorflow_tagging/tensorflow_tagging.py b/Python Server/app/image_tagging/tensorflow_tagging/tensorflow_tagging.py
@@ -0,0 +1,42 @@
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+
+import numpy as np
+
+from tensorflow.python.keras.applications.inception_v3 import InceptionV3, preprocess_input, decode_predictions
+from tensorflow.python.keras.preprocessing import image
+import os
+from utils import download_image_from_url, createFolder
+
+
+class TensorFlowTag:
+
+    def __init__(self, delete_image_after_tagging=True):
+        self.delete_after_tagging = delete_image_after_tagging
+
+        createFolder("./image_tagging/temp_image_folder")
+
+    def get_tags(self, image_url):
+        model = InceptionV3(weights='imagenet')
+        file = download_image_from_url(image_url)
+        img = image.load_img(file, target_size=(299, 299))
+        x = image.img_to_array(img)
+        x = np.expand_dims(x, axis=0)
+        x = preprocess_input(x)
+
+        preds = model.predict(x)
+        prediction_decoded = decode_predictions(preds, top=5)[0]
+
+        print('Predicted:', prediction_decoded)
+
+        tags = []
+        for element in prediction_decoded:
+            tags.append(element[1])
+
+        if self.delete_after_tagging:
+            os.remove(file)
+
+        str1 = ','.join(str(e) for e in tags)
+
+        return str1
diff --git a/Python Server/app/requirements.txt b/Python Server/app/requirements.txt