import numpy as np
import pandas as pd
import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt
from scipy import stats

np.random.seed(42)
genes = ['MAP6', 'PPARGC1A', 'RELN', 'HCN1', 'SLC16A2', 'IDH2', 'MCU', 'TREM2', 'APOE', 'APP', 'MAPT', 'PSEN1', 'CD33', 'CLU', 'BIN1', 'GFAP', 'AQP4', 'SLC17A7', 'GAD1', 'MBP']
cell_types = ['Microglia', 'Astrocytes', 'Exc_Neurons', 'Inh_Neurons', 'Oligodendrocytes']
conditions = ['Control', 'AD']
n_samples = 50

expression_data = {}
for gene in genes:
    expression_data[gene] = {}
    for ct in cell_types:
        ctrl = np.random.lognormal(mean=2.0, sigma=0.5, size=n_samples)
        if gene in ['TREM2', 'CD33', 'C1QA'] and ct == 'Microglia':
            ad = ctrl * np.random.lognormal(mean=0.8, sigma=0.3, size=n_samples)
        elif gene in ['GFAP'] and ct == 'Astrocytes':
            ad = ctrl * np.random.lognormal(mean=0.6, sigma=0.3, size=n_samples)
        elif gene in ['SLC17A7', 'SYP'] and 'Neuron' in ct:
            ad = ctrl * np.random.lognormal(mean=-0.3, sigma=0.2, size=n_samples)
        elif gene in ['MBP'] and ct == 'Oligodendrocytes':
            ad = ctrl * np.random.lognormal(mean=-0.2, sigma=0.2, size=n_samples)
        else:
            ad = ctrl * np.random.lognormal(mean=0.1, sigma=0.3, size=n_samples)
        expression_data[gene][ct] = {'Control': ctrl, 'AD': ad}

print(f"Generated expression data for {len(genes)} genes x {len(cell_types)} cell types")
print(f"Samples per condition: {n_samples}")

# --- Expression Heatmap: Log2 Fold Change ---
log2fc = np.zeros((len(genes), len(cell_types)))
pvalues = np.zeros((len(genes), len(cell_types)))

for i, gene in enumerate(genes):
    for j, ct in enumerate(cell_types):
        ctrl = expression_data[gene][ct]['Control']
        ad = expression_data[gene][ct]['AD']
        log2fc[i, j] = np.log2(np.mean(ad) / np.mean(ctrl))
        _, pvalues[i, j] = stats.mannwhitneyu(ctrl, ad, alternative='two-sided')

fig, ax = plt.subplots(figsize=(10, 8))
im = ax.imshow(log2fc, cmap='RdBu_r', vmin=-2, vmax=2, aspect='auto')

ax.set_xticks(range(len(cell_types)))
ax.set_xticklabels([ct.replace('_', ' ') for ct in cell_types], rotation=45, ha='right', fontsize=9)
ax.set_yticks(range(len(genes)))
ax.set_yticklabels(genes, fontsize=9)

for i in range(len(genes)):
    for j in range(len(cell_types)):
        sig = ''
        if pvalues[i, j] < 0.001: sig = '***'
        elif pvalues[i, j] < 0.01: sig = '**'
        elif pvalues[i, j] < 0.05: sig = '*'
        color = 'white' if abs(log2fc[i, j]) > 1 else 'black'
        ax.text(j, i, f'{log2fc[i,j]:.2f}\n{sig}', ha='center', va='center',
                fontsize=7, color=color)

cbar = plt.colorbar(im, ax=ax, label='Log2 Fold Change (AD/Control)')
ax.set_title('Differential Gene Expression: AD vs Control by Cell Type', fontsize=12, fontweight='bold')
fig.patch.set_facecolor('#0a0a14')
ax.set_facecolor('#151525')
ax.tick_params(colors='#e0e0e0')
ax.title.set_color('#4fc3f7')
cbar.ax.yaxis.set_tick_params(color='#e0e0e0')
cbar.ax.yaxis.label.set_color('#e0e0e0')
plt.setp(plt.getp(cbar.ax.axes, 'yticklabels'), color='#e0e0e0')
plt.tight_layout()
plt.show()

print(f"\nSignificant changes (p < 0.05): {int(np.sum(pvalues < 0.05))} / {pvalues.size}")

# --- Volcano Plot: Microglia Expression Changes ---
fig, ax = plt.subplots(figsize=(10, 7))
fig.patch.set_facecolor('#0a0a14')
ax.set_facecolor('#151525')

fc_vals = []
pv_vals = []
gene_labels = []

for gene in genes:
    ctrl = expression_data[gene]['Microglia']['Control']
    ad = expression_data[gene]['Microglia']['AD']
    fc = np.log2(np.mean(ad) / np.mean(ctrl))
    _, pv = stats.mannwhitneyu(ctrl, ad, alternative='two-sided')
    fc_vals.append(fc)
    pv_vals.append(-np.log10(max(pv, 1e-300)))
    gene_labels.append(gene)

fc_vals = np.array(fc_vals)
pv_vals = np.array(pv_vals)

colors = []
for fc, pv in zip(fc_vals, pv_vals):
    if pv > -np.log10(0.05) and abs(fc) > 0.5:
        colors.append('#ef5350' if fc > 0 else '#4fc3f7')
    else:
        colors.append('#555555')

ax.scatter(fc_vals, pv_vals, c=colors, s=80, alpha=0.8, edgecolors='white', linewidths=0.5)

for i, (fc, pv, label) in enumerate(zip(fc_vals, pv_vals, gene_labels)):
    if pv > -np.log10(0.05) and abs(fc) > 0.3:
        ax.annotate(label, (fc, pv), fontsize=8, color='#e0e0e0',
                   xytext=(5, 5), textcoords='offset points')

ax.axhline(-np.log10(0.05), color='#ffd54f', linestyle='--', alpha=0.5, label='p=0.05')
ax.axvline(0.5, color='#81c784', linestyle='--', alpha=0.3)
ax.axvline(-0.5, color='#81c784', linestyle='--', alpha=0.3)
ax.set_xlabel('Log2 Fold Change (AD/Control)', color='#e0e0e0', fontsize=11)
ax.set_ylabel('-Log10(p-value)', color='#e0e0e0', fontsize=11)
ax.set_title('Volcano Plot: Microglia Gene Expression in AD', fontsize=12, fontweight='bold', color='#4fc3f7')
ax.tick_params(colors='#e0e0e0')
ax.legend(facecolor='#151525', edgecolor='#333', labelcolor='#e0e0e0')
plt.tight_layout()
plt.show()

sig_up = sum(1 for fc, pv in zip(fc_vals, pv_vals) if pv > -np.log10(0.05) and fc > 0.5)
sig_down = sum(1 for fc, pv in zip(fc_vals, pv_vals) if pv > -np.log10(0.05) and fc < -0.5)
print(f"\nSignificantly upregulated: {sig_up}")
print(f"Significantly downregulated: {sig_down}")

# --- Statistical Tests ---
print("=" * 70)
print("STATISTICAL ANALYSIS SUMMARY")
print("=" * 70)

results = []
for gene in genes[:10]:
    for ct in cell_types:
        ctrl = expression_data[gene][ct]['Control']
        ad = expression_data[gene][ct]['AD']
        stat_mw, p_mw = stats.mannwhitneyu(ctrl, ad, alternative='two-sided')
        pooled_std = np.sqrt((np.std(ctrl)**2 + np.std(ad)**2) / 2)
        cohens_d = (np.mean(ad) - np.mean(ctrl)) / pooled_std if pooled_std > 0 else 0

        if p_mw < 0.05:
            results.append({
                'Gene': gene,
                'Cell_Type': ct.replace('_', ' '),
                'Log2FC': np.log2(np.mean(ad) / np.mean(ctrl)),
                'P_value': p_mw,
                'Cohens_d': cohens_d,
                'Effect': 'Large' if abs(cohens_d) > 0.8 else ('Medium' if abs(cohens_d) > 0.5 else 'Small'),
            })

stats_df = pd.DataFrame(results)
if len(stats_df) > 0:
    stats_df = stats_df.sort_values('P_value')
    print(f"\nSignificant results (p < 0.05): {len(stats_df)} gene-cell type pairs\n")
    print(stats_df.head(15).to_string(index=False))

    print("\n" + "=" * 70)
    print("ONE-WAY ANOVA: Expression Variation Across Cell Types (AD samples)")
    print("=" * 70)
    for gene in genes[:5]:
        groups = [expression_data[gene][ct]['AD'] for ct in cell_types]
        f_stat, p_anova = stats.f_oneway(*groups)
        sig = "***" if p_anova < 0.001 else ("**" if p_anova < 0.01 else ("*" if p_anova < 0.05 else "ns"))
        print(f"  {gene:15s}  F={f_stat:8.2f}  p={p_anova:.2e}  {sig}")
else:
    print("No significant results found at p < 0.05")

# --- Pathway Enrichment Analysis ---
from scipy.stats import hypergeom

target_genes = ['MAP6', 'PPARGC1A', 'RELN', 'HCN1', 'SLC16A2', 'IDH2', 'MCU']

pathways = {
    'Amyloid Processing': ['APP', 'PSEN1', 'PSEN2', 'BACE1', 'ADAM10', 'APH1A'],
    'Tau Phosphorylation': ['MAPT', 'GSK3B', 'CDK5', 'DYRK1A', 'PP2A', 'MARK4'],
    'Microglial Activation': ['TREM2', 'CD33', 'TYROBP', 'SPI1', 'C1QA', 'C3'],
    'Neuroinflammation': ['IL1B', 'TNF', 'IL6', 'NFKB1', 'NLRP3', 'CXCL10'],
    'Lipid Metabolism': ['APOE', 'CLU', 'ABCA7', 'ABCA1', 'LDLR', 'LRP1'],
    'Synaptic Function': ['SYP', 'SLC17A7', 'SNAP25', 'DLG4', 'GRIA1', 'GRIN2B'],
    'Mitochondrial Function': ['PINK1', 'PRKN', 'SIRT3', 'TFAM', 'PPARGC1A', 'DNM1L'],
    'Autophagy-Lysosome': ['LAMP1', 'SQSTM1', 'BECN1', 'ATG5', 'TFEB', 'GRN'],
    'Oxidative Stress': ['SOD1', 'SOD2', 'GPX4', 'NRF2', 'HMOX1', 'CAT'],
    'Calcium Signaling': ['CALM1', 'CAMK2A', 'ITPR1', 'RYR2', 'ATP2A2', 'CACNA1C'],
    'Ferroptosis': ['ACSL4', 'GPX4', 'SLC7A11', 'LPCAT3', 'TFRC', 'FTH1'],
    'Epigenetic Regulation': ['HDAC1', 'DNMT1', 'TET2', 'KDM6A', 'EZH2', 'SIRT1'],
}

total_genes = 20000
query_size = len(target_genes)

enrichment_results = []
for pathway_name, pathway_genes in pathways.items():
    overlap = set(target_genes) & set(pathway_genes)
    if len(overlap) > 0 or np.random.random() < 0.3:
        k = len(overlap)
        M = total_genes
        n = len(pathway_genes)
        N = query_size
        pval = hypergeom.sf(k - 1, M, n, N) if k > 0 else 1.0
        fold_enrichment = (k / max(N, 1)) / (n / M) if n > 0 else 0
        enrichment_results.append({
            'Pathway': pathway_name,
            'Genes_in_Pathway': n,
            'Overlap': k,
            'Overlap_Genes': ', '.join(overlap) if overlap else '-',
            'Fold_Enrichment': fold_enrichment,
            'P_value': pval,
        })

enrich_df = pd.DataFrame(enrichment_results).sort_values('P_value')
enrich_df['Significant'] = enrich_df['P_value'] < 0.05
print("Pathway Enrichment Results:")
print(enrich_df[['Pathway', 'Overlap', 'Fold_Enrichment', 'P_value', 'Significant']].to_string(index=False))