etrotta commited on
Commit
b6be18b
·
1 Parent(s): a3e2f28

Change Database port to work on HuggingFace Spaces properly

Browse files
Files changed (2) hide show
  1. config.py +11 -3
  2. database.py +6 -1
config.py CHANGED
@@ -6,11 +6,19 @@ qdrant_api_key = os.getenv('QDRANT_API_KEY')
6
  description = """This is a Kanji image search demo. Draw or upload an image of an individual Kanji character."""
7
 
8
  article = """
 
 
 
 
 
 
 
 
9
  ### About this project
10
 
11
- You can find the source code as well as more information in https://github.com/etrotta/kanji_lookup
12
 
13
- It uses the "kha-white/manga-ocr-base" ViT Encoder model to create embeddings, then uses a vector database (qdrant) to find similar characters.
14
 
15
- The vector database has been populated with over 10k characters from [The KANJIDIC project](https://www.edrdg.org/wiki/index.php/KANJIDIC_Project), each rendered in multiple fonts downloaded from Google Fonts
16
  """
 
6
  description = """This is a Kanji image search demo. Draw or upload an image of an individual Kanji character."""
7
 
8
  article = """
9
+ ### Getting better results
10
+
11
+ Try different brush sizes.
12
+ Try to draw it centered in the middle of the canvas, both horizontally and vertically.
13
+ You may want to try using an external tool to draw then import a file.
14
+
15
+ The results is sorted by estimated distance from the input, but will rarely give the exact Kanji you are searching for as the first result
16
+
17
  ### About this project
18
 
19
+ It uses the "kha-white/manga-ocr-base" Vision Transformer Encoder model to create embeddings, then uses a vector database (qdrant) to find similar characters.
20
 
21
+ You can find the code used to create the embeddings as well as more information in https://github.com/etrotta/kanji_lookup
22
 
23
+ The database has been populated with over 10000 characters from [The KANJIDIC project](https://www.edrdg.org/wiki/index.php/KANJIDIC_Project), each rendered in multiple fonts downloaded from Google Fonts
24
  """
database.py CHANGED
@@ -4,7 +4,12 @@ from qdrant_client import QdrantClient, models
4
 
5
  from config import qdrant_location, qdrant_api_key
6
 
7
- qdrant = QdrantClient(qdrant_location, api_key=qdrant_api_key)
 
 
 
 
 
8
 
9
  def search_vector(query_vector: torch.Tensor, limit: int=20) -> list[models.ScoredPoint]:
10
  hits = qdrant.search(
 
4
 
5
  from config import qdrant_location, qdrant_api_key
6
 
7
+ qdrant = QdrantClient(
8
+ qdrant_location,
9
+ api_key=qdrant_api_key,
10
+ port=443,
11
+ timeout=30,
12
+ )
13
 
14
  def search_vector(query_vector: torch.Tensor, limit: int=20) -> list[models.ScoredPoint]:
15
  hits = qdrant.search(