1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
use clap::{Parser, ValueEnum};
#[derive(Clone, Copy, Debug, Default, Eq, PartialEq, ValueEnum)]
pub enum SeedProfile {
/// Server-side SQL (`generate_series` / recursive CTE). Best for 100k+ rows.
#[default]
Fast,
/// Realistic names, JSON payloads, and lorem text. Postgres uses COPY + workers.
Realistic,
/// Original v1 multi-row INSERT from Rust (slow; backward compatibility).
Insert,
}
fn default_workers() -> usize {
std::thread::available_parallelism()
.map(|n| n.get().clamp(1, 8))
.unwrap_or(4)
}
#[derive(Parser)]
#[command(
name = "seed",
about = "Generate test data for rivet (v2: fast SQL profile + COPY bulk load)"
)]
pub struct Args {
/// Target database: postgres, mysql, or both
#[arg(short, long, default_value = "both")]
pub target: String,
/// Load strategy: `fast` (SQL-side generation) or `realistic` (Rust + COPY on PG)
#[arg(long, value_enum, default_value_t = SeedProfile::Fast)]
pub profile: SeedProfile,
/// Parallel COPY/insert workers (realistic Postgres orders/events/page_views)
#[arg(long, default_value_t = default_workers())]
pub workers: usize,
/// Number of users to generate
#[arg(long, default_value = "100000")]
pub users: usize,
/// Average orders per user (fast profile uses exact count; realistic uses Poisson)
#[arg(long, default_value = "10")]
pub orders_per_user: usize,
/// Average events per user (fast profile uses exact count; realistic uses Poisson)
#[arg(long, default_value = "50")]
pub events_per_user: usize,
/// Number of page_views to generate (wide table, degraded scenario)
#[arg(long, default_value = "2000000")]
pub page_views: usize,
/// Number of content_items to generate (heavy text, worst case for memory)
#[arg(long, default_value = "200000")]
pub content_items: usize,
/// PostgreSQL connection URL
#[arg(long, default_value = "postgresql://rivet:rivet@localhost:5432/rivet")]
pub pg_url: String,
/// MySQL connection URL
#[arg(long, default_value = "mysql://rivet:rivet@localhost:3306/rivet")]
pub mysql_url: String,
/// Batch size for realistic INSERT path (Postgres realistic uses COPY instead)
#[arg(long, default_value = "1000")]
pub batch_size: usize,
/// Fill `orders_sparse` with few rows and huge gaps in `id` (for chunked / sparse-key demos)
#[arg(long)]
pub sparse_chunk_demo: bool,
/// Only create (if needed), truncate, and fill `orders_sparse` — skip users/orders/events/page_views/content
#[arg(long)]
pub only_sparse_chunk_demo: bool,
/// Number of rows in `orders_sparse` (ids: 1, 1+gap, 1+2·gap, …). Use ≥3 to see sparse min..max vs row count
#[arg(long, default_value = "3")]
pub sparse_chunk_rows: usize,
/// Gap between consecutive sparse ids (wider gap ⇒ more empty chunk windows for the same chunk_size)
#[arg(long, default_value = "2000000")]
pub sparse_chunk_id_gap: i64,
/// Rows per INSERT for `orders_sparse` (keep ≤ few thousand if max_allowed_packet / statement size is tight)
#[arg(long, default_value = "5000")]
pub sparse_chunk_batch_size: usize,
/// Number of rows in `orders_coalesce` (composite-cursor demo with NULL updated_at)
#[arg(long, default_value = "2000")]
pub coalesce_rows: usize,
/// Fraction of `orders_coalesce` rows with NULL `updated_at` (range 0.0..1.0)
#[arg(long, default_value = "0.35")]
pub coalesce_null_ratio: f64,
}
impl Args {
pub fn planned_orders(&self) -> usize {
self.users.saturating_mul(self.orders_per_user)
}
pub fn planned_events(&self) -> usize {
self.users.saturating_mul(self.events_per_user)
}
}