Major update. Support for 15 LLMs, World Flora Online taxonomy validation, geolocation, 2 OCR methods, significant UI changes, stability improvements, consistent JSON parsing

app.py

@@ -1,5 +1,5 @@
 import streamlit as st
-import yaml, os, json, random, time, re, torch, random, warnings, shutil, sys
+import yaml, os, json, random, time, re, torch, random, warnings, shutil, sys, glob
 import seaborn as sns
 import plotly.graph_objs as go
 from PIL import Image
@@ -14,7 +14,7 @@ from vouchervision.vouchervision_main import voucher_vision
 from vouchervision.general_utils import test_GPU, get_cfg_from_full_path, summarize_expense_report, validate_dir
 from vouchervision.model_maps import ModelMaps
 from vouchervision.API_validation import APIvalidation
-from vouchervision.utils_hf import setup_streamlit_config, save_uploaded_file, save_uploaded_local
+from vouchervision.utils_hf import setup_streamlit_config, save_uploaded_file, save_uploaded_local, save_uploaded_file_local
 from vouchervision.data_project import convert_pdf_to_jpg
 from vouchervision.utils_LLM import check_system_gpus
 
@@ -42,7 +42,7 @@ if 'config' not in st.session_state:
     st.session_state.config, st.session_state.dir_home = build_VV_config(loaded_cfg=None)
     setup_streamlit_config(st.session_state.dir_home)
 
-# st.session_state['is_hf'] = True
+st.session_state['is_hf'] = True
 
 ########################################################################################################
 ###  Global constants                                                                               ####
@@ -273,7 +273,7 @@ def content_input_images(col_left, col_right):
         if st.session_state.is_hf:
             if uploaded_files:
                 # Clear input image gallery and input list
-                clear_image_gallery()
+                clear_image_uploads()
 
                 for uploaded_file in uploaded_files:
                     # Determine the file type
@@ -336,6 +336,45 @@ def content_input_images(col_left, col_right):
                 pass
             # elif st.session_state['input_list_small'] and (st.session_state['dir_images_local_TEMP'] != st.session_state.config['leafmachine']['project']['dir_images_local']):
             elif (st.session_state['dir_images_local_TEMP'] != st.session_state.config['leafmachine']['project']['dir_images_local']):
+                has_pdf = False
+                clear_image_uploads()
+
+                for input_file in os.listdir(st.session_state.config['leafmachine']['project']['dir_images_local']):
+                    if input_file.split('.')[1].lower() in ['jpg','jpeg']:
+                        pass
+                    elif input_file.split('.')[1].lower() in ['pdf',]:
+                        has_pdf = True
+                        # Handle PDF files
+                        file_path = save_uploaded_file_local(st.session_state.config['leafmachine']['project']['dir_images_local'], st.session_state['dir_uploaded_images'], input_file)
+                        # Convert each page of the PDF to an image
+                        n_pages = convert_pdf_to_jpg(file_path, st.session_state['dir_uploaded_images'], dpi=200)#st.session_state.config['leafmachine']['project']['dir_images_local'])
+                        
+                        
+                        # pdf_files_pattern = os.path.join(st.session_state['dir_uploaded_images'], '*.pdf')
+                        # for pdf_file in glob.glob(pdf_files_pattern):
+                        #     os.remove(pdf_file)
+
+                        # # Update the input list for each page image
+                        # converted_files = os.listdir(st.session_state['dir_uploaded_images'])
+                        # for file_name in converted_files:   
+                        #     if file_name.lower().endswith('.jpg'):
+                        #         jpg_file_path = os.path.join(st.session_state['dir_uploaded_images'], file_name)
+                        #         st.session_state['input_list'].append(jpg_file_path)
+
+                        #         # Optionally, create a thumbnail for the gallery
+                        #         img = Image.open(jpg_file_path)
+                        #         img.thumbnail((GALLERY_IMAGE_SIZE, GALLERY_IMAGE_SIZE), Image.Resampling.LANCZOS)
+                        #         file_path_small = save_uploaded_file_local(st.session_state['dir_uploaded_images'], st.session_state['dir_uploaded_images_small'], file_name, img)
+                        #         st.session_state['input_list_small'].append(file_path_small)
+                        
+                        # st.session_state.config['leafmachine']['project']['dir_images_local'] = st.session_state['dir_uploaded_images']
+                        
+                    else:
+                        pass
+                        # st.warning("Inputs must be '.PDF' or '.jpg' or '.jpeg'")
+                if has_pdf:
+                    st.session_state.config['leafmachine']['project']['dir_images_local'] = st.session_state['dir_uploaded_images']
+
                 dir_images_local = st.session_state.config['leafmachine']['project']['dir_images_local']
                 count_n_imgs = list_jpg_files(dir_images_local)
                 st.session_state['processing_add_on'] = count_n_imgs
@@ -412,6 +451,10 @@ def delete_directory(dir_path):
 
 
 def clear_image_gallery():
+    delete_directory(st.session_state['dir_uploaded_images_small'])
+    validate_dir(st.session_state['dir_uploaded_images_small'])
+
+def clear_image_uploads():
     delete_directory(st.session_state['dir_uploaded_images'])
     delete_directory(st.session_state['dir_uploaded_images_small'])
     validate_dir(st.session_state['dir_uploaded_images'])
@@ -423,7 +466,7 @@ def use_test_image():
     st.session_state.config['leafmachine']['project']['dir_images_local'] = os.path.join(st.session_state.dir_home,'demo','demo_images')
     n_images = len([f for f in os.listdir(st.session_state.config['leafmachine']['project']['dir_images_local']) if os.path.isfile(os.path.join(st.session_state.config['leafmachine']['project']['dir_images_local'], f))])
     st.session_state['processing_add_on'] = n_images
-    clear_image_gallery()
+    clear_image_uploads()
     st.session_state['uploader_idk'] += 1
     for file in os.listdir(st.session_state.config['leafmachine']['project']['dir_images_local']):
         file_path = save_uploaded_file(os.path.join(st.session_state.dir_home,'demo','demo_images'), file)

vouchervision/OCR_google_cloud_vision.py

@@ -144,6 +144,9 @@ class OCREngine:
 
     def init_gemini_vision(self):
         pass
+
+    def init_gpt4_vision(self):
+        pass
             
 
     def detect_text_craft(self):

vouchervision/utils_hf.py

@@ -6,7 +6,7 @@ import base64
 from PIL import Image
 from PIL import Image
 from io import BytesIO
-from shutil import copyfileobj
+from shutil import copyfileobj, copyfile
 
 # from vouchervision.general_utils import get_cfg_from_full_path
 
@@ -37,6 +37,37 @@ def setup_streamlit_config(dir_home):
         f.write(config_content.strip())
 
 
+def save_uploaded_file_local(directory_in, directory_out, img_file_name, image=None):
+    if not os.path.exists(directory_out):
+        os.makedirs(directory_out)
+
+    # Assuming img_file_name includes the extension
+    img_file_base, img_file_ext = os.path.splitext(img_file_name)
+    
+    full_path_out = os.path.join(directory_out, img_file_name)
+    full_path_in = os.path.join(directory_in, img_file_name)
+
+    # Check if the file extension is .pdf (or add other conditions for different file types)
+    if img_file_ext.lower() == '.pdf':
+        # Copy the file from the input directory to the output directory
+        copyfile(full_path_in, full_path_out)
+        return full_path_out
+    else:
+        if image is None:
+            try:
+                with Image.open(full_path_in) as image:
+                    image.save(full_path_out, "JPEG")
+                # Return the full path of the saved image
+                return full_path_out
+            except:
+                pass
+        else:
+            try:
+                image.save(full_path_out, "JPEG")
+                return full_path_out
+            except:
+                pass
+            
 
 def save_uploaded_file(directory, img_file, image=None):
     if not os.path.exists(directory):