import streamlit as st import json from history_tree import HistoryTree from utils import save_json, DEFAULTS from streamlit_agraph import agraph, Node, Edge, Config def render_timeline_wip(data, file_path): tree_data = data.get("history_tree", {}) if not tree_data: st.info("No history timeline exists.") return htree = HistoryTree(tree_data) # ========================================== # 1. BUILD GRAPH DATA (NODES & EDGES) # ========================================== nodes = [] edges = [] sorted_nodes = sorted(htree.nodes.values(), key=lambda x: x["timestamp"]) for n in sorted_nodes: nid = n["id"] note = n.get('note', 'Step') short_note = (note[:15] + '..') if len(note) > 15 else note # Default Styles color = "#ffffff" border = "#666666" # Highlight Head (Current Pointer) if nid == htree.head_id: color = "#fff6cd" border = "#eebb00" # Highlight Tips (Ends of branches) if nid in htree.branches.values(): if color == "#ffffff": color = "#e6ffe6" border = "#44aa44" nodes.append(Node( id=nid, label=f"{short_note}\n({nid[:4]})", size=25, shape="box", color=color, borderWidth=1, borderColor=border, font={'color': 'black', 'face': 'Arial', 'size': 14} )) if n["parent"] and n["parent"] in htree.nodes: edges.append(Edge( source=n["parent"], target=nid, color="#aaaaaa", type="STRAIGHT" )) # ========================================== # 2. RENDER INTERACTIVE GRAPH # ========================================== config = Config( width="100%", height="500px", directed=True, physics=False, hierarchical=True, layout={ "hierarchical": { "enabled": True, "levelSeparation": 150, "nodeSpacing": 100, "treeSpacing": 100, "direction": "LR", "sortMethod": "directed" } } ) st.subheader("✨ Interactive Timeline") st.caption("Click any node to preview its settings and compare changes.") clicked_node_id = agraph(nodes=nodes, edges=edges, config=config) st.markdown("---") # ========================================== # 3. INSPECTION PANEL # ========================================== target_node_id = clicked_node_id if clicked_node_id else htree.head_id if target_node_id and target_node_id in htree.nodes: selected_node = htree.nodes[target_node_id] node_data = selected_node["data"] c_h1, c_h2 = st.columns([3, 1]) c_h1.markdown(f"### 🔎 Inspecting: {selected_node.get('note', 'Step')}") c_h1.caption(f"ID: {target_node_id}") # --- A. COMPARE CHANGES (DIFF) --- with st.expander(f"📊 Compare Changes", expanded=False): diffs = [] all_keys = set(data.keys()) | set(node_data.keys()) # Keys to ignore in diff view to reduce noise ignore_keys = { "history_tree", "prompt_history", "batch_data", "ui_reset_token", "sequence_number", "input_a_frames", "input_b_frames" } for k in all_keys: if k in ignore_keys: continue val_now = data.get(k, "") val_then = node_data.get(k, "") # Convert to string and strip whitespace for clean comparison str_now = str(val_now).strip() str_then = str(val_then).strip() if str_now != str_then: # Fuzzy match for numbers (ignore 1 vs 1.0) try: f_now = float(str_now) f_then = float(str_then) if abs(f_now - f_then) < 0.001: continue except ValueError: pass diffs.append((k, str_now, str_then)) if not diffs: st.caption("✅ Identical to current state") else: for k, v_now, v_then in diffs: dc1, dc2, dc3 = st.columns([1, 2, 2]) dc1.markdown(f"**{k}**") dc2.markdown(f"🔴 `{str(v_now)[:30]}`") dc3.markdown(f"🟢 `{str(v_then)[:30]}`") # --- B. RESTORE ACTION --- with c_h2: st.write(""); st.write("") if st.button("⏪ Restore This Version", type="primary", use_container_width=True): data.update(node_data) htree.head_id = target_node_id data["history_tree"] = htree.to_dict() save_json(file_path, data) st.session_state.ui_reset_token += 1 label = f"{selected_node.get('note')} ({target_node_id[:4]})" st.session_state.restored_indicator = label st.toast(f"Restored {target_node_id}!", icon="🔄") st.rerun() # --- C. SNAPSHOT PREVIEW (READ ONLY FORM) --- st.markdown("#### 📄 Snapshot Preview") # 1. Prompts (Handles both old 'positive_prompt' and new 'general_prompt' keys) p_col1, p_col2 = st.columns(2) with p_col1: val_p = node_data.get("positive_prompt", "") or node_data.get("general_prompt", "") st.text_area("Positive Prompt", value=val_p, height=100, disabled=True, key="prev_pp") with p_col2: val_n = node_data.get("negative_prompt", "") or node_data.get("general_negative", "") st.text_area("Negative Prompt", value=val_n, height=100, disabled=True, key="prev_np") # 2. Key Settings s_col1, s_col2, s_col3, s_col4 = st.columns(4) s_col1.text_input("Seed", value=str(node_data.get("seed", "")), disabled=True, key="prev_seed") s_col2.text_input("Steps", value=str(node_data.get("steps", "")), disabled=True, key="prev_steps") s_col3.text_input("CFG", value=str(node_data.get("cfg", "")), disabled=True, key="prev_cfg") s_col4.text_input("Denoise", value=str(node_data.get("denoise", "")), disabled=True, key="prev_den") # 3. Models m_col1, m_col2 = st.columns(2) m_col1.text_input("Checkpoint", value=node_data.get("model_name", ""), disabled=True, key="prev_ckpt") m_col2.text_input("VAE", value=node_data.get("vae_name", ""), disabled=True, key="prev_vae") # 4. LoRAs with st.expander("💊 LoRA Configuration"): l1, l2, l3 = st.columns(3) with l1: st.text_input("L1 Name", value=node_data.get("lora 1 high", ""), disabled=True, key="prev_l1h") st.text_input("L1 Str", value=str(node_data.get("lora 1 low", "")), disabled=True, key="prev_l1l") with l2: st.text_input("L2 Name", value=node_data.get("lora 2 high", ""), disabled=True, key="prev_l2h") st.text_input("L2 Str", value=str(node_data.get("lora 2 low", "")), disabled=True, key="prev_l2l") with l3: st.text_input("L3 Name", value=node_data.get("lora 3 high", ""), disabled=True, key="prev_l3h") st.text_input("L3 Str", value=str(node_data.get("lora 3 low", "")), disabled=True, key="prev_l3l")