Update to use spacy

README.md

@@ -11,27 +11,16 @@ queries.
 
 ## Installation
 ```
-pip install git+https://github.com/TakoData/QueryFilter.git@0.1.2
+pip install git+https://github.com/TakoData/QueryFilter.git@0.1.4
 ```
-With embeddings support, which pulls in the large `sentence-transformers` package:
-```
-pip install "git+https://github.com/TakoData/QueryFilter.git@0.1.2#egg=tako-query-filter[embeddings]"
-```
-
 ## Usage
 
-### Prerequisites
-- Get access to Tako Hugging Face repositories
-- Install and init `git-lfs`
-- Log into Hugging Face using `huggingface-cli login`
-
-### Examples
 See the [demo notebook](demo.ipynb) for a more interactive example.
 
 ```
 from tako_query_filter.filter import TakoQueryFilter
 
-query_filter = TakoQueryFilter.load_from_hf()
+query_filter = TakoQueryFilter.load_with_keywords()
 
 queries = [
     "aapl vs nvda",
@@ -42,7 +31,6 @@ queries = [
 
 query_filter.predict(queries)
 ```
-You can also use the `force_download` flag to download the latest version of the models from HuggingFace.
 
 ## License
 

demo.ipynb

@@ -1,15 +1,5 @@
 {
  "cells": [
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "# Make sure you are logged in to Hugging Face\n",
-    "!huggingface-cli login --token \"hf_my-hf-token\""
-   ]
-  },
   {
    "cell_type": "code",
    "execution_count": null,
@@ -23,19 +13,13 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 5,
    "metadata": {},
    "outputs": [],
    "source": [
     "from tako_query_filter.filter import TakoQueryFilter\n",
     "\n",
-    "query_filter = TakoQueryFilter.load_from_hf(force_download=True)\n",
-    "# Alternatively, you can load from local paths\n",
-    "# query_filter = TakoQueryFilter.load_from_local(\n",
-    "#     topic_model_path=\"local-scikit-model-path/topic_model.pkl\",\n",
-    "#     spacy_model_path=\"local-spacy-model-path\",\n",
-    "#     keywords_path=\"local-scikit-model-path/topic_model.pkl\"\n",
-    "# )\n"
+    "query_filter = TakoQueryFilter.load_with_keywords()\n"
    ]
   },
   {
@@ -53,10 +37,11 @@
     "    \"trump vs. harris\",\n",
     "    \"algae\",\n",
     "    \"san francisco game this week\",\n",
-    "    \"what college did michael phelps go to\"\n",
+    "    \"what college did michael phelps go to\",\n",
     "    \"election 2024\",\n",
     "    \"election results\",\n",
     "    \"presidential election polls 2024\",\n",
+    "    \"what books do you recommend\",\n",
     "]\n",
     "\n",
     "preds = query_filter.predict(queries)\n",