from huggingface_hub import hf_hub_download
import json
import pandas as pd
from pathlib import Path
import gradio as gr

REPO_ID = "ashutoshzade/EIRKG"
FILENAME = "EIRKG_v1.1_full.json"


def load_dataset():

    try:

        print("Downloading EIRKG dataset...")

        path = hf_hub_download(
            repo_id=REPO_ID,
            repo_type="dataset",
            filename=FILENAME
        )

        print("Success!")

        if FILENAME.endswith(".json"):
            with open(path,"r",encoding="utf-8") as f:
                return json.load(f)

        if FILENAME.endswith(".jsonl"):
            data=[]
            with open(path,"r",encoding="utf-8") as f:
                for line in f:
                    data.append(json.loads(line))
            return data

        if FILENAME.endswith(".csv"):
            return pd.read_csv(path).to_dict("records")

    except Exception as e:

        print(e)

        return []

DATA = load_dataset()

print(len(DATA))

# ---------------------
# Build searchable text
# ---------------------

SEARCH = []

for row in DATA:

    combined = ""

    for value in row.values():
        combined += str(value).lower() + " "

    SEARCH.append(combined)

# ---------------------
# Search function
# ---------------------

def query_eirkg(question):

    q = question.lower()

    scores = []

    words = q.split()

    for idx, text in enumerate(SEARCH):

        score = 0

        for word in words:
            score += text.count(word)

        scores.append((score, idx))

    scores.sort(reverse=True)

    output = ""

    count = 0

    for score, idx in scores:

        if score == 0:
            continue

        row = DATA[idx]

        output += f"# Match {count+1}\n\n"

        for k, v in row.items():

            if str(v).strip() != "":
                output += f"**{k}**: {v}\n\n"

        output += "---\n\n"

        count += 1

        if count >= 5:
            break

    if output == "":
        output = "No matching EIRKG entries found."

    return output

# ---------------------
# Gradio UI
# ---------------------

DESCRIPTION = """

# Engineering and Innovation Reasoning Knowledge Graph (EIRKG)

# Engineering and Innovation Corpus to demostate how cross domain solutions can be formed with innovative LLM techniques. 

Copyright (C) 2026 Ashutosh Zade

AGPL 3.0

Ask engineering questions to explore the EIRKG dataset.

Examples:

• battery recycling

• electric locomotive

• autonomous mining

• digital twin

• biomimicry

• carbon neutral transportation

"""

demo = gr.Interface(

    fn=query_eirkg,

    inputs=gr.Textbox(
        label="Engineering Question",
        placeholder="Ask about batteries, locomotives, mining..."
    ),

    outputs=gr.Markdown(label="EIRKG Results"),

    title="EIRKG Explorer",

    description=DESCRIPTION

)

demo.launch()