path: root/internal/parser/tabular_csv_test.go

package parser

import (
	"context"
	"fmt"
	"strings"
	"testing"
	"time"

	"epimetheus/internal/metrics"
)

func TestTabularCSVParser_Parse(t *testing.T) {
	tests := []struct {
		name          string
		input         string
		metricName    string
		expectedCount int
		wantErr       bool
	}{
		{
			name: "simple tabular CSV with numeric and text columns",
			input: `responsecode,httpmethod,user,totaltime
200,GET,alice,50.5
404,POST,bob,100.2
500,GET,charlie,75.0`,
			metricName:    "test_metric",
			expectedCount: 6, // 3 rows * 2 numeric columns (responsecode, totaltime)
			wantErr:       false,
		},
		{
			name: "CSV with mixed data types",
			input: `col1,col2,col3
1,text,3.14
2,more,2.71
3,data,1.41`,
			metricName:    "mixed_metric",
			expectedCount: 6, // 3 rows * 2 numeric columns (col1, col3)
			wantErr:       false,
		},
		{
			name: "CSV with whitespace",
			input: `  col1  ,  col2  ,  col3  
  1  ,  value2  ,  3  
  4  ,  value5  ,  6  `,
			metricName:    "whitespace_metric",
			expectedCount: 4, // 2 rows * 2 numeric columns (col1, col3)
			wantErr:       false,
		},
		{
			name: "CSV with comments",
			input: `# This is a comment
col1,col2,col3
# Another comment
1,value2,3`,
			metricName:    "comment_metric",
			expectedCount: 2, // 1 row * 2 numeric columns (col1, col3)
			wantErr:       false,
		},
		{
			name:          "empty CSV",
			input:         "",
			metricName:    "empty_metric",
			expectedCount: 0,
			wantErr:       true, // No header row
		},
		{
			name: "header only",
			input: `col1,col2,col3
`,
			metricName:    "header_only",
			expectedCount: 0,
			wantErr:       false,
		},
		{
			name: "mismatched columns - skipped",
			input: `col1,col2,col3
1,value2,3
value4,value5
6,value7,8`,
			metricName:    "mismatched_metric",
			expectedCount: 4, // 2 matching rows * 2 numeric columns
			wantErr:       false,
		},
	}

	for _, tt := range tests {
		t.Run(tt.name, func(t *testing.T) {
			ctx := context.Background()
			timestamp := time.Now()
			parser := NewTabularCSVParser(tt.metricName, timestamp, []string{}) // No DNS resolution

			reader := strings.NewReader(tt.input)
			samples, err := parser.Parse(ctx, reader)

			if (err != nil) != tt.wantErr {
				t.Errorf("Parse() error = %v, wantErr %v", err, tt.wantErr)
				return
			}

			if err == nil && len(samples) != tt.expectedCount {
				t.Errorf("Parse() got %d samples, want %d", len(samples), tt.expectedCount)
			}

			// Verify all samples have the correct base metric name and timestamp
			for _, sample := range samples {
				if !strings.HasPrefix(sample.MetricName, tt.metricName+"_") {
					t.Errorf("Sample metric name = %s, want prefix %s_", sample.MetricName, tt.metricName)
				}
				if !sample.Timestamp.Equal(timestamp) {
					t.Errorf("Sample timestamp = %v, want %v", sample.Timestamp, timestamp)
				}
			}
		})
	}
}

func TestTabularCSVParser_Labels(t *testing.T) {
	ctx := context.Background()
	timestamp := time.Now()
	input := `responsecode,httpmethod,user,totaltime
200,GET,alice,50.5
404,POST,bob,100.2`

	parser := NewTabularCSVParser("test_metric", timestamp, []string{})
	reader := strings.NewReader(input)
	samples, err := parser.Parse(ctx, reader)

	if err != nil {
		t.Fatalf("Parse() error = %v", err)
	}

	// 2 rows * 2 numeric columns (responsecode, totaltime) = 4 samples
	if len(samples) != 4 {
		t.Fatalf("Expected 4 samples, got %d", len(samples))
	}

	// Find the responsecode and totaltime metrics for first row
	var responsecodeMetric, totaltimeMetric *metrics.Sample
	for i := range samples {
		if samples[i].Labels["httpmethod"] == "GET" && samples[i].Labels["user"] == "alice" {
			if strings.HasSuffix(samples[i].MetricName, "_responsecode") {
				responsecodeMetric = &samples[i]
			} else if strings.HasSuffix(samples[i].MetricName, "_totaltime") {
				totaltimeMetric = &samples[i]
			}
		}
	}

	if responsecodeMetric == nil || totaltimeMetric == nil {
		t.Fatalf("Could not find expected metrics")
	}

	// Check responsecode metric
	if responsecodeMetric.Value != 200 {
		t.Errorf("responsecode value = %f, want 200", responsecodeMetric.Value)
	}
	expectedLabels := map[string]string{
		"httpmethod": "GET",
		"user":       "alice",
	}
	for key, expectedValue := range expectedLabels {
		if actualValue, ok := responsecodeMetric.Labels[key]; !ok {
			t.Errorf("responsecode metric missing label %s", key)
		} else if actualValue != expectedValue {
			t.Errorf("responsecode metric label %s = %s, want %s", key, actualValue, expectedValue)
		}
	}

	// Check totaltime metric
	if totaltimeMetric.Value != 50.5 {
		t.Errorf("totaltime value = %f, want 50.5", totaltimeMetric.Value)
	}
	for key, expectedValue := range expectedLabels {
		if actualValue, ok := totaltimeMetric.Labels[key]; !ok {
			t.Errorf("totaltime metric missing label %s", key)
		} else if actualValue != expectedValue {
			t.Errorf("totaltime metric label %s = %s, want %s", key, actualValue, expectedValue)
		}
	}
}

func TestTabularCSVParser_ContextCancellation(t *testing.T) {
	ctx, cancel := context.WithCancel(context.Background())
	cancel() // Cancel immediately

	timestamp := time.Now()
	input := `col1,col2,col3
value1,value2,value3
value4,value5,value6`

	parser := NewTabularCSVParser("test_metric", timestamp, []string{})
	reader := strings.NewReader(input)
	_, err := parser.Parse(ctx, reader)

	if err != context.Canceled {
		t.Errorf("Expected context.Canceled error, got %v", err)
	}
}

func TestTabularCSVParser_LargeFile(t *testing.T) {
	ctx := context.Background()
	timestamp := time.Now()

	// Generate a CSV with 1000 rows, 1 numeric column
	var builder strings.Builder
	builder.WriteString("col1,col2,col3\n")
	for i := 0; i < 1000; i++ {
		builder.WriteString(fmt.Sprintf("%d,value2,value3\n", i))
	}

	parser := NewTabularCSVParser("large_metric", timestamp, []string{})
	reader := strings.NewReader(builder.String())
	samples, err := parser.Parse(ctx, reader)

	if err != nil {
		t.Fatalf("Parse() error = %v", err)
	}

	// 1000 rows * 1 numeric column = 1000 samples
	if len(samples) != 1000 {
		t.Errorf("Expected 1000 samples, got %d", len(samples))
	}
}

func TestSanitizeLabelName(t *testing.T) {
	tests := []struct {
		input    string
		expected string
	}{
		{"avg(totaltime)", "avg_totaltime"},
		{"sum(rcv)", "sum_rcv"},
		{"response-code", "response_code"},
		{"valid_label", "valid_label"},
		{"ValidLabel123", "ValidLabel123"},
		{"123invalid", "label_123invalid"},
		{"label__with___underscores", "label_with_underscores"},
		{"_leading", "leading"},
		{"trailing_", "trailing"},
		{"special!@#$chars", "special_chars"},
		{"", "unknown"},
		{"___", "unknown"},
		{"http.method", "http_method"},
		{"status_code", "status_code"},
	}

	for _, tt := range tests {
		t.Run(tt.input, func(t *testing.T) {
			result := sanitizeLabelName(tt.input)
			if result != tt.expected {
				t.Errorf("sanitizeLabelName(%q) = %q, want %q", tt.input, result, tt.expected)
			}
		})
	}
}

func TestTabularCSVParser_WithSpecialCharacterHeaders(t *testing.T) {
	ctx := context.Background()
	timestamp := time.Now()
	input := `avg(totaltime),sum(rcv),response-code,http.method
50.5,1102,200,GET
100.2,2204,404,POST`

	parser := NewTabularCSVParser("special_metric", timestamp, []string{})
	reader := strings.NewReader(input)
	samples, err := parser.Parse(ctx, reader)

	if err != nil {
		t.Fatalf("Parse() error = %v", err)
	}

	// 2 rows * 3 numeric columns (avg(totaltime), sum(rcv), response-code) = 6 samples
	if len(samples) != 6 {
		t.Fatalf("Expected 6 samples, got %d", len(samples))
	}

	// Check that metric names are sanitized
	expectedMetrics := []string{"special_metric_avg_totaltime", "special_metric_sum_rcv", "special_metric_response_code"}
	foundMetrics := make(map[string]bool)
	for _, sample := range samples {
		foundMetrics[sample.MetricName] = true
	}
	for _, expected := range expectedMetrics {
		if !foundMetrics[expected] {
			t.Errorf("Missing expected metric %s", expected)
		}
	}

	// Check that string column becomes label (http.method -> http_method)
	for _, sample := range samples {
		if _, ok := sample.Labels["http_method"]; !ok {
			t.Errorf("Sample missing sanitized label http_method")
		}
	}

	// Check values are correct for first row
	for _, sample := range samples {
		if sample.Labels["http_method"] == "GET" {
			if strings.HasSuffix(sample.MetricName, "_avg_totaltime") {
				if sample.Value != 50.5 {
					t.Errorf("avg_totaltime value = %f, want 50.5", sample.Value)
				}
			} else if strings.HasSuffix(sample.MetricName, "_sum_rcv") {
				if sample.Value != 1102 {
					t.Errorf("sum_rcv value = %f, want 1102", sample.Value)
				}
			} else if strings.HasSuffix(sample.MetricName, "_response_code") {
				if sample.Value != 200 {
					t.Errorf("response_code value = %f, want 200", sample.Value)
				}
			}
		}
	}
}

func TestTabularCSVParser_DNSResolution(t *testing.T) {
	ctx := context.Background()
	timestamp := time.Now()
	
	// Test with localhost IP which should resolve on most systems
	input := `ip,responsecode,count
127.0.0.1,200,100`

	parser := NewTabularCSVParser("test_metric", timestamp, []string{"ip"})
	reader := strings.NewReader(input)
	samples, err := parser.Parse(ctx, reader)

	if err != nil {
		t.Fatalf("Parse() error = %v", err)
	}

	// Should have 1 sample (count metric)
	if len(samples) != 1 {
		t.Fatalf("Expected 1 sample, got %d", len(samples))
	}

	// The ip label should either be resolved to a hostname or remain as IP if DNS failed
	ipLabel := samples[0].Labels["ip"]
	if ipLabel == "" {
		t.Error("ip label is empty")
	}
	
	// If resolution succeeded, it should not be the original IP
	// If it failed, it should still be the IP
	t.Logf("IP label value: %s (original: 127.0.0.1)", ipLabel)
}

func TestTabularCSVParser_DNSResolutionDisabled(t *testing.T) {
	ctx := context.Background()
	timestamp := time.Now()
	
	input := `ip,responsecode,count
192.168.1.1,200,100`

	// Pass empty slice - no DNS resolution
	parser := NewTabularCSVParser("test_metric", timestamp, []string{})
	reader := strings.NewReader(input)
	samples, err := parser.Parse(ctx, reader)

	if err != nil {
		t.Fatalf("Parse() error = %v", err)
	}

	if len(samples) != 1 {
		t.Fatalf("Expected 1 sample, got %d", len(samples))
	}

	// IP should remain unchanged
	ipLabel := samples[0].Labels["ip"]
	if ipLabel != "192.168.1.1" {
		t.Errorf("IP label = %s, want 192.168.1.1 (DNS resolution should be disabled)", ipLabel)
	}
}

func TestTabularCSVParser_DNSResolutionMultipleLabels(t *testing.T) {
	ctx := context.Background()
	timestamp := time.Now()
	
	input := `source_ip,dest_ip,count
127.0.0.1,192.168.1.1,100`

	// Resolve both source_ip and dest_ip
	parser := NewTabularCSVParser("test_metric", timestamp, []string{"source_ip", "dest_ip"})
	reader := strings.NewReader(input)
	samples, err := parser.Parse(ctx, reader)

	if err != nil {
		t.Fatalf("Parse() error = %v", err)
	}

	if len(samples) != 1 {
		t.Fatalf("Expected 1 sample, got %d", len(samples))
	}

	// Both labels should be present
	if _, ok := samples[0].Labels["source_ip"]; !ok {
		t.Error("source_ip label missing")
	}
	if _, ok := samples[0].Labels["dest_ip"]; !ok {
		t.Error("dest_ip label missing")
	}
	
	t.Logf("source_ip: %s", samples[0].Labels["source_ip"])
	t.Logf("dest_ip: %s", samples[0].Labels["dest_ip"])
}

func TestTabularCSVParser_DNSResolutionNonIPValue(t *testing.T) {
	ctx := context.Background()
	timestamp := time.Now()
	
	// ip column contains non-IP value
	input := `ip,responsecode,count
not-an-ip,200,100`

	parser := NewTabularCSVParser("test_metric", timestamp, []string{"ip"})
	reader := strings.NewReader(input)
	samples, err := parser.Parse(ctx, reader)

	if err != nil {
		t.Fatalf("Parse() error = %v", err)
	}

	if len(samples) != 1 {
		t.Fatalf("Expected 1 sample, got %d", len(samples))
	}

	// Non-IP value should remain unchanged
	ipLabel := samples[0].Labels["ip"]
	if ipLabel != "not-an-ip" {
		t.Errorf("IP label = %s, want not-an-ip (should not resolve non-IP values)", ipLabel)
	}
}

func TestTabularCSVParser_DNSResolutionCaching(t *testing.T) {
	ctx := context.Background()
	timestamp := time.Now()
	
	// Multiple rows with same IP
	input := `ip,responsecode,count
127.0.0.1,200,100
127.0.0.1,404,50
127.0.0.1,500,5`

	parser := NewTabularCSVParser("test_metric", timestamp, []string{"ip"})
	reader := strings.NewReader(input)
	samples, err := parser.Parse(ctx, reader)

	if err != nil {
		t.Fatalf("Parse() error = %v", err)
	}

	// Should have 3 samples
	if len(samples) != 3 {
		t.Fatalf("Expected 3 samples, got %d", len(samples))
	}

	// All should have same resolved IP (cached result)
	firstIP := samples[0].Labels["ip"]
	for i, sample := range samples {
		if sample.Labels["ip"] != firstIP {
			t.Errorf("Sample %d has different IP label: %s vs %s (caching issue)", i, sample.Labels["ip"], firstIP)
		}
	}
	
	// Check cache size
	if parser.resolver.GetCacheSize() != 1 {
		t.Errorf("Expected cache size 1 (one unique IP), got %d", parser.resolver.GetCacheSize())
	}
}