import gradio as gr
import pandas as pd

from huggingface_hub import hf_hub_url, cached_download
from gensim.models import KeyedVectors

# Setup model
url = hf_hub_url(repo_id="map222/recipe-spice-model", filename="recipe_w2v.gensim")
cached_download(url)
recipe_w2v = KeyedVectors.load(cached_download(url) )

'''
url = hf_hub_url(repo_id="map222/recipe-spice-model", filename="ingredient_count.json")
cached_download(url)
ingredient_count = json.load(cached_download(url) )

url = hf_hub_url(repo_id="map222/recipe-spice-model", filename="recipe_NER.tsv")
cached_download(url)
recipe_NER = pd.read_csv(cached_download(url), sep='\t' )

'''

def get_fusion_ingredients(ingredient: str,
                           recipe_model, #gensim model
                           recipes, #iterable of recipes
                           ingredient_count: dict,
                           max_candidates = 20,
                           min_occurence_factor = 100 # minimum number of recipes an ingredient has to be in
                           ):

    ingredient_recipes = recipes.loc[recipes.apply(lambda row: ingredient in row)]
  
    ingredient_candidates = recipe_model.most_similar(ingredient, topn=50) # get top similar ingredients
    candidate_names = list(zip(*ingredient_candidates))[0]
    pruned_candidates = [candidate for candidate in candidate_names if ingredient not in candidate][:max_candidates] # clean up candidates to remove duplicates (e.g. "gala apple")
    cooccurrence_counts = calc_cooccurrence(ingredient, candidate_names, ingredient_recipes) # get counts for normalization
    # final score for sorting: similarity / how often co-occur / total occurences
    min_occurences = max(cooccurrence_counts.values()) / min_occurence_factor
    print(min_occurences)
    freq_norm_candidates = {candidate[0]: candidate[1] / (cooccurrence_counts[candidate[0]]+1) / ingredient_count[candidate[0]] for candidate in ingredient_candidates if candidate[0] in pruned_candidates and cooccurrence_counts[candidate[0]] > min_occurences}
    top_candidates = sorted([(k,v) for k,v in freq_norm_candidates.items()], key=lambda x: x[1])[-5:]
    return top_candidates, cooccurrence_counts, ingredient_candidates # return multiple for debugging

def helper_func(text):
    #a,b,c = get_fusion_ingredients(x, recipe_w2v, ingredient_count, recipe_NER)
    nearest = recipe_w2v.wv.most_similar(text)
    return pd.DataFrame(nearest)

app = gr.Blocks()

with app:
    gr.Markdown("# Recipe Spice")
    with gr.Row():
        text_in = gr.Textbox(lines=1, placeholder="Ingredient", label="Search Query")
        
        
    with gr.Row():
        df_out = gr.Dataframe(interactive=False)
    with gr.Row():
        gr.Markdown(
            """
            #### Project Description
            This finds cool new recipe stuff
            """
        )
    text_in.submit(helper_func, inputs=[text_in], outputs=[df_out])

app.launch()