app.py

import spaces                        # for ZeroGPU support
import gradio as gr
import pandas as pd
import numpy as np
import torch
from threading import Thread
from transformers import (
    AutoModelForCausalLM,
    AutoTokenizer,
    AutoProcessor,
    TextIteratorStreamer
)

# ─── MODEL SETUP ────────────────────────────────────────────────────────────────
MODEL_NAME = "bytedance-research/ChatTS-14B"

tokenizer = AutoTokenizer.from_pretrained(
    MODEL_NAME, trust_remote_code=True
)
processor = AutoProcessor.from_pretrained(
    MODEL_NAME, trust_remote_code=True, tokenizer=tokenizer
)
model = AutoModelForCausalLM.from_pretrained(
    MODEL_NAME,
    trust_remote_code=True,
    device_map="auto",
    torch_dtype=torch.float16
)
model.eval()

# ─── VISUAL THEME (Dark) ───────────────────────────────────────────────────────
CSS = """
.gradio-container {
  box-shadow: none !important;
  border: none !important;
}
:root{
  --bg:#0b1020; --bg-2:#10172e; --panel:#141b34; --panel-2:#0f172a;
  --border:#263154; --fg:#e7ecf5; --muted:#b9c2d3; --placeholder:#7c8195;
}
/* App background */
.gradio-container{background: linear-gradient(180deg,var(--bg),var(--bg) 40%, var(--bg-2)); color: var(--fg);}
/* Generic block panels */
.gradio-container .block, .gradio-container .form, .gradio-container .padded{
  background: linear-gradient(180deg,var(--panel),var(--panel-2));
  border: 1px solid var(--border);
  border-radius: 14px;
}
/* Inputs */
input[type="text"], textarea, select, .gr-textbox, .gr-dropdown, .gr-textbox textarea{
  background: #0f1630 !important; color: var(--fg) !important;
  border: 1px solid var(--border) !important; border-radius: 12px !important;
}
.gr-textbox textarea::placeholder{ color: var(--placeholder) !important; }
label, .svelte-1gfkn6j, .svelte-i3tvor, .label{ color: var(--muted) !important; }

/* Upload panel */
button.svelte-1x5qevo{ background:#0f1630 !important; color:var(--muted) !important; border-color:var(--border) !important; }
.svelte-12ioyct{ color: var(--muted) !important; }

/* Line plot container */
.vega-embed, canvas.marks{ background:#0b1020 !important; border-radius: 10px; }

/* Buttons base */
.gr-button{ border-radius:14px !important; border:1px solid var(--border) !important; color:#0b1020 !important; font-weight:700; }
.gr-button:hover{ transform: translateY(-1px); box-shadow:0 6px 18px rgba(0,0,0,.35); }
.gr-button-primary{ background: linear-gradient(180deg,#8ab4ff,#7aa2ff) !important; }
.gr-button-secondary{ background: linear-gradient(180deg,#a78bfa,#7aa2ff) !important; color:#0b1020 !important; }
.btn-chip{ padding:12px 16px !important; }

/* Example cards row */
.examples-row{ display:grid; grid-template-columns:repeat(auto-fit,minmax(180px,1fr)); gap:12px; margin:12px 0 8px; }
.cardbtn{ border:1px solid var(--border) !important; color:#0b1020 !important; }
.grad-blue{ background: linear-gradient(180deg,#93c5fd,#60a5fa) !important; }
.grad-purple{ background: linear-gradient(180deg,#c4b5fd,#a78bfa) !important; }
.grad-green{ background: linear-gradient(180deg,#bbf7d0,#34d399) !important; }
.grad-amber{ background: linear-gradient(180deg,#fde68a,#f59e0b) !important; }
.grad-rose{ background: linear-gradient(180deg,#fecaca,#fb7185) !important; }
.cardbtn:hover{ filter:brightness(1.02); transform:translateY(-1px); }

/* Subtle spacing fixes */
.gap, .padded{ padding:12px !important; }
"""

# ─── HELPERS ───────────────────────────────────────────────────────────────────
def create_default_timeseries():
    x = np.arange(256)
    ts1 = np.sin(x / 10) * 5.0
    ts1[103:] -= 10.0
    ts2 = x * 0.01
    ts2[100] += 10.0
    return pd.DataFrame({"TS1": ts1, "TS2": ts2})

def process_csv_file(csv_file):
    if csv_file is None:
        return None, "No file uploaded"
    try:
        df = pd.read_csv(csv_file.name)
        df.columns = [str(c).strip() for c in df.columns]
        df = df.loc[:, [c for c in df.columns if c]]
        df = df.dropna(axis=1, how="all")
        if df.shape[1] == 0:
            return None, "No valid time-series columns found."
        if df.shape[1] > 15:
            return None, f"Too many series ({df.shape[1]}). Max allowed = 15."
        ts_names = []
        ts_list = []
        for name in df.columns:
            series = df[name]
            if not pd.api.types.is_float_dtype(series):
                series = pd.to_numeric(series, errors='coerce')
            last_valid = series.last_valid_index()
            if last_valid is None:
                continue
            trimmed = series.loc[:last_valid].to_numpy(dtype=np.float32)
            L = trimmed.shape[0]
            if L < 64 or L > 1024:
                return None, f"Series '{name}' length {L} invalid. Must be 64 to 1024."
            ts_names.append(name); ts_list.append(trimmed)
        if not ts_list:
            return None, "All time series are empty after trimming NaNs."
        return df, f"Loaded {len(ts_names)} series: {', '.join(ts_names)}"
    except Exception as e:
        return None, f"Error processing file: {str(e)}"

def preview_csv(csv_file, use_default):
    if csv_file is None:
        return gr.LinePlot(value=pd.DataFrame()), "Please upload a CSV file first", gr.Dropdown(), False
    df, message = process_csv_file(csv_file)
    if df is None:
        return gr.LinePlot(value=pd.DataFrame()), message, gr.Dropdown(), False
    choices = list(df.columns)
    first = choices[0]
    df_idx = df.copy(); df_idx["_i"] = np.arange(len(df[first].values))
    plot = gr.LinePlot(df_idx, x="_i", y=first, title=f"Time Series: {first}")
    dropdown = gr.Dropdown(choices=choices, value=first, label="Select a Column to Visualize")
    return plot, message, dropdown, False

def clear_csv():
    return gr.LinePlot(value=pd.DataFrame()), "Cleared.", gr.Dropdown()

def update_plot(csv_file, selected_column, use_default_state):
    if (csv_file is None and not use_default_state) or selected_column is None:
        return gr.LinePlot(value=pd.DataFrame())
    if csv_file is None and use_default_state:
        df = create_default_timeseries()
    else:
        df, _ = process_csv_file(csv_file)
    if df is None:
        return gr.LinePlot(value=pd.DataFrame())
    df_idx = df.copy(); df_idx["_i"] = np.arange(len(df[selected_column].values))
    return gr.LinePlot(df_idx, x="_i", y=selected_column, title=f"Time Series: {selected_column}")

def initialize_interface():
    df = create_default_timeseries()
    choices = list(df.columns); first = choices[0]
    df_idx = df.copy(); df_idx["_i"] = np.arange(len(df[first].values))
    plot = gr.LinePlot(df_idx, x="_i", y=first, title=f"Time Series: {first}")
    dropdown = gr.Dropdown(choices=choices, value=first, label="Select a Column to Visualize")
    msg = "Using default time series (TS1 and TS2). Select a series from the dropdown for visualization."
    return plot, msg, dropdown, True

# ─── INFERENCE ─────────────────────────────────────────────────────────────────
@spaces.GPU
def infer_chatts_stream(prompt: str, csv_file, use_default):
    if not prompt.strip():
        yield "Please enter a prompt"
        return
    if csv_file is None and use_default:
        df = create_default_timeseries()
        error_msg = None
    else:
        df, error_msg = process_csv_file(csv_file)
    if df is None:
        yield "Please upload a CSV file first or the file contains errors"
        return
    try:
        ts_names, ts_list = [], []
        for name in df.columns:
            series = df[name]; last_valid = series.last_valid_index()
            if last_valid is not None:
                trimmed = series.loc[:last_valid].to_numpy(dtype=np.float32)
                ts_names.append(name); ts_list.append(trimmed)
        if not ts_list:
            yield "No valid time series data found. Please upload time series first."
            return
        clean_prompt = prompt.replace("<ts>", "").replace("<ts/>", "")
        prefix = f"I have {len(ts_list)} time series:\n"
        for name, arr in zip(ts_names, ts_list):
            prefix += f"The {name} is of length {len(arr)}: <ts><ts/>\n"
        full_prompt = (
            "<|im_start|>system\nYou are a helpful assistant. Your name is ChatTS. "
            "You can analyze time series data and provide insights. If user asks who you are, you should give your name and capabilities "
            "in the language of the prompt. If no time series are provided, you should say 'I cannot answer this question as you haven't provide the timeseries I need' "
            "in the language of the prompt. Always check if the user has provided at least one time series data before answering."
            "<|im_end|><|im_start|>user\n"
            f"{prefix}{clean_prompt} Please output a step-by-step analysis about the time series attributes that mentioned in the question first, "
            "and then give a detailed result about this question. Always remember to carefully double check the values before answer the results."
            "<|im_end|><|im_start|>assistant\n"
        )
        inputs = processor(text=[full_prompt], timeseries=ts_list, padding=True, return_tensors="pt")
        inputs = {k: v.to(model.device) for k, v in inputs.items()}
        streamer = TextIteratorStreamer(tokenizer, timeout=10., skip_prompt=True, skip_special_tokens=True)
        inputs.update({"max_new_tokens": 512, "streamer": streamer, "temperature": 0.3})
        thread = Thread(target=model.generate, kwargs=inputs); thread.start()
        out = ""
        for new_text in streamer:
            out += new_text
            yield out
    except Exception as e:
        yield f"Error during inference: {str(e)}"

# ─── EXAMPLES (semantic names) ─────────────────────────────────────────────────
EXAMPLE_PROMPTS = {
    "🔍 Detect Spikes": "Identify all spikes in each series and report timestamps and magnitudes. Explain briefly.",
    "📈 Trend & Seasonality": "Describe the trend and seasonality for the provided time series TS1.",
    "🔗 Compare Metrics": "Compare the two series (TS1 and TS2). Are there lagged correlations? Estimate the lag and correlation strength.",
    "⚡ Local Change Analysis": "Find intervals with >10% rise or drop relative to the prior 20 points for TS1. Return intervals and reasons.",
    "📊 Correlation Strength": "Quantify Pearson correlation between each pair of series (TS1 and TS2) and highlight the strongest positive/negative pairs."
}

# ─── UI ────────────────────────────────────────────────────────────────────────
with gr.Blocks(title="ChatTS Demo", css=CSS) as demo:
    use_default_state = gr.State(value=True)

    gr.HTML("<h3 style='text-align:center; color:#e8ebfa; font-size:20px; font-weight:700; margin-top:8px;'>💡 Click an example below and click Run ChatTS to try different questions with default time series or upload your own CSV file.</h3>")

    with gr.Row(elem_classes="examples-row"):
        btns = {}
        grads = ["grad-blue","grad-purple","grad-green","grad-amber","grad-rose"]
        for (title, prompt), grad in zip(EXAMPLE_PROMPTS.items(), grads):
            btns[title] = gr.Button(title, elem_classes=f"cardbtn btn-chip {grad}")

    with gr.Row():
        with gr.Column(scale=1):
            upload = gr.File(label="", file_types=[".csv"], type="filepath")
            prompt_input = gr.Textbox(
                lines=6,
                placeholder="Enter your question here...",
                label="Analysis Prompt",
                value="Find the maximum and minimum values in each series and comment on them."
            )
            run_btn = gr.Button("Run ChatTS", variant="primary")
        with gr.Column(scale=2):
            series_selector = gr.Dropdown(label="Select a Column to Visualize", choices=[], value=None)
            plot_out = gr.LinePlot(value=pd.DataFrame(), label="Time Series Visualization")
            file_status = gr.Textbox(label="File Status", interactive=False, lines=2)

    text_out = gr.Textbox(lines=10, label="ChatTS Analysis Results", interactive=False)

    demo.load(fn=initialize_interface, outputs=[plot_out, file_status, series_selector, use_default_state])

    for title, prompt in EXAMPLE_PROMPTS.items():
        btns[title].click(fn=lambda p=prompt: p, outputs=prompt_input)

    upload.upload(fn=preview_csv, inputs=[upload, use_default_state],
                  outputs=[plot_out, file_status, series_selector, use_default_state])
    upload.clear(fn=clear_csv, outputs=[plot_out, file_status, series_selector])
    series_selector.change(fn=update_plot, inputs=[upload, series_selector, use_default_state], outputs=[plot_out])
    run_btn.click(fn=infer_chatts_stream, inputs=[prompt_input, upload, use_default_state], outputs=[text_out])

if __name__ == '__main__':
    demo.launch()