Storing demo data in home folder instead of local

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "vectory"
-version = "0.1.4"
+version = "0.1.5"
 description = "Streamline the benchmark and experimentation process of your models that rely on generating embeddings"
 authors = ["Pento <hello@pento.ai>"]
 readme = "README.md"
@@ -32,7 +32,7 @@ classifiers= [
  "Topic :: Software Development :: Libraries",
 ]
 
-include = ["docker-compose.yml", "Dockerfile", ".dockerignore", "troubleshooting.md"]
+include = ["docker-compose.yml", "Dockerfile", ".dockerignore", "TROUBLESHOOTING.md"]
 
 [tool.poetry.scripts]
 vectory = "vectory.cli:app"

vectory/__init__.py

@@ -8,6 +8,6 @@
 from .demo import download_demo_data, prepare_demo_data # noqa
 from .es.client import ElasticKNNClient # noqa
 from .es.utils import load_csv_with_headers, load_embeddings_from_numpy # noqa
-from .experiments import Experiment
+from .experiments import Experiment # noqa
 from .indices import delete_index, list_indices, load_index, match_query # noqa
 from .spaces import EmbeddingSpace, compare_embedding_spaces # noqa

vectory/cli.py

@@ -24,6 +24,7 @@
 from vectory.experiments import Experiment
 from vectory.indices import delete_index, list_indices, load_index
 from vectory.spaces import EmbeddingSpace, compare_embedding_spaces
+from vectory.utils import get_vectory_dir
 from vectory.visualization.run import run_streamlit
 
 create_db_tables()
@@ -77,7 +78,7 @@ def demo(
  ),
  ),
  data_path: Path = typer.Option(
- "data/demo",
+ get_vectory_dir() / "demo",
  "--data-path",
  help="Path to the demo files",
  file_okay=False,

vectory/db/models.py

@@ -4,13 +4,13 @@
 
 from peewee import CharField, DateTimeField, IntegerField, Model
 from playhouse.sqlite_ext import ForeignKeyField, JSONField, SqliteExtDatabase
+from vectory.utils import get_vectory_dir
 
-default_folder_path = os.path.join(os.path.expanduser("~"), ".vectory")
-db_path = os.path.join(default_folder_path, "main.db")
+DB_PATH = get_vectory_dir() / "main.db"
 
 
 database = SqliteExtDatabase(
- db_path,
+ DB_PATH,
  pragmas={
  "journal_mode": "off",
  "synchronous": 0,
@@ -216,8 +216,8 @@ def get_knn(embedding_space_name: str, metric: str):
 
 
 def create_db_tables():
- if not os.path.exists(default_folder_path):
- os.makedirs(default_folder_path)
+ if not DB_PATH.parent.exists():
+ os.makedirs(DB_PATH.parent, exist_ok=True)
  with database:
  database.create_tables(
  [

vectory/demo.py

@@ -4,7 +4,7 @@
 from pathlib import Path
 from typing import List, Union
 
-import requests
+import requests # type: ignore
 import typer
 from vectory.datasets import Dataset
 from vectory.db.models import EmbeddingSpaceModel
@@ -30,7 +30,7 @@ def download_demo_data(
 ) -> None:
 
  if not os.path.isdir(data_path):
- os.makedirs(data_path)
+ os.makedirs(data_path, exist_ok=True)
 
  base_url = "https://github.com/pentoai/vectory/releases/download/v0.1.1/"
 
@@ -71,13 +71,7 @@ def prepare_demo_data(
 
  dataset = Dataset.get_or_create(
  name=dataset_name,
- csv_path=str(
- Path(__file__).parent.parent
- / "data"
- / "demo"
- / dataset_name
- / f"{dataset_name}-data.csv"
- ),
+ csv_path=str(Path(data_path) / f"{dataset_name}-data.csv"),
  id_field="_idx",
  )
  typer.secho(f"Dataset {dataset_name} created", fg="yellow")

vectory/utils.py

@@ -1,5 +1,11 @@
+from pathlib import Path
+
 from coolname import generate
 
 
 def generate_name(name: str, words=3) -> str:
  return "-".join([*generate(words), name])
+
+
+def get_vectory_dir() -> Path:
+ return Path.home() / ".vectory"

vectory/visualization/main.py

@@ -80,7 +80,7 @@ def selection(dataset):
  selected_emb_space, model=model, similarity=similarity
  )
 
- except:
+ except Exception:
  st.warning(
  "There isn't any loaded index from the embedding space "
  + f"'{selected_emb_space}', with model: '{model}' "
@@ -450,7 +450,11 @@ def main():
  most_similar_indices_2, scores_2 = calculate_indices(
  selected_vector, index_2
  )
- intersection = set(most_similar_indices_1) & set(most_similar_indices_1)
+
+ # intersection = (
+ # set(most_similar_indices_1) & set(most_similar_indices_1)
+ # )
+
  col1, col2 = st.columns(2)
  if df_1 is not None:
 

vectory/visualization/run.py

@@ -1,4 +1,3 @@
-import os
 from pathlib import Path
 
 from streamlit import config