image upload, rough ui but working

app.py

@@ -124,6 +124,18 @@ def image_retrieval_from_image(state, selected_locally):
     return features_to_gallery(embeddings[selected])
 
 
+def query_uploaded_image(uploaded_image):
+    image = preprocess(uploaded_image)
+    image_batch = torch.tensor(np.stack([image])).to(device)
+    with torch.no_grad():
+        image_features = model.encode_image(image_batch).float()
+        image_features = image_features.cpu().numpy()
+        assert len(image_features) == 1
+        image_features = image_features[0]
+        assert len(image_features) == d
+    return features_to_gallery(image_features)
+
+
 def show_folder(state, selected_locally):
     if state is None or len(state) == 0:
         return [], []
@@ -155,6 +167,10 @@ with gr.Blocks(css="footer {visibility: hidden}") as demo:
         image_query_button = gr.Button("Show similar")
         selected = gr.Number(0, show_label=False, visible=False)
 
+    uploaded_image = gr.Image(tool="select", type="pil")
+
+    query_uploaded_image_button = gr.Button("Show similiar to uploaded")
+
     gallery = gr.Gallery(label="Images", show_label=False, elem_id="gallery"
         ).style(columns=4, container=False)
 
@@ -164,6 +180,7 @@ with gr.Blocks(css="footer {visibility: hidden}") as demo:
     text_query_button.click(image_retrieval_from_text, [text], [gallery, state])
     image_query_button.click(image_retrieval_from_image, [state, selected], [gallery, state])
     show_folder_button.click(show_folder, [state, selected], [gallery, state])
+    query_uploaded_image_button.click(query_uploaded_image, [uploaded_image], [gallery, state])
 
     def get_select_index(evt: gr.SelectData, state):
         selected_locally = evt.index

readme.sh

@@ -26,8 +26,8 @@ python convert.py sample_fbi_s1e1.pkl
 
 ssh -p 2820 hexagon.renyi.hu
 cd ai-shared/daniel/sameenergy
-lftp -p 2167 gw.pioneer.hu
-# manually provide username, password
+lftp -p 2167 -u d.varga gw.pioneer.hu
+# manually provide password
 cd store/05_Photos
 # promising directories:
 ls 02_LOCATION\ PHOTOS 05_TO_LOCATION_PHOTOS PhotoLibrary Tünde
@@ -39,6 +39,12 @@ ctrl-d
 
 # scp'd files to buda
 cd /data/daniel/sameenergy/
+
+# how many bytes, as a check?
+find 02_LOCATION_PHOTOS -type f -exec stat --format="%s" {} \; | awk '{total += $1} END {print total}'
+# -> 141,133,402,112 that's 141GB. in 197108 files, not including directories.
+# on the Pioneer server this was 141,131,778,304 bytes in 196916 files, 6446 directories, good enough.
+
 find 02_LOCATION_PHOTOS -type f > raw_files
 cat raw_files | grep -i "jpg\|jpeg$" > jpg_files
 
@@ -55,6 +61,9 @@ nohup bash create_embeddings.sh &
 bash hashes.sh
 # takes jpg_files and outputs md5sums
 
+# ad hoc sample from duplicates:
+cat md5sums | awk 'BEGIN{FS="  "} { if ($1 in m) { print $1 "\t" $2 "\t" m[$1] } ; m[$1] = $2 }' | awk '(NR%4000==0)' | cut -f2-
+
 python convert.py 02_LOCATION_PHOTOS.pkl
 # -> creates float16 02_LOCATION_PHOTOS.f16.pkl
 mv md5sums 02_LOCATION_PHOTOS.f16.md5sums