@kensio/part-factory-test-data 1.3.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/.claude-plugin/plugin.json +14 -0
- package/README.md +60 -0
- package/package.json +33 -0
- package/skills/part-factory-test-data/SKILL.md +235 -0
|
@@ -0,0 +1,14 @@
|
|
|
1
|
+
{
|
|
2
|
+
"$schema": "https://anthropic.com/claude-code/plugin.schema.json",
|
|
3
|
+
"name": "part-factory-test-data",
|
|
4
|
+
"version": "1.3.0",
|
|
5
|
+
"description": "How to build test data with @kensio/part-factory: keeping in the factory everything a test does not care about, passing dependencies at call time so factories stay independent and shareable, and choosing between its static, dynamic, variant and mapped factories.",
|
|
6
|
+
"author": {
|
|
7
|
+
"name": "Kensio Software",
|
|
8
|
+
"email": "hugh@kensiosoftware.co.uk"
|
|
9
|
+
},
|
|
10
|
+
"homepage": "https://kensio.ai",
|
|
11
|
+
"repository": "https://github.com/KensioSoftware/kensio.ai",
|
|
12
|
+
"license": "Apache-2.0",
|
|
13
|
+
"keywords": ["testing", "test-data", "factory", "part-factory", "faker"]
|
|
14
|
+
}
|
package/README.md
ADDED
|
@@ -0,0 +1,60 @@
|
|
|
1
|
+
# @kensio/part-factory-test-data
|
|
2
|
+
|
|
3
|
+
A Claude Code skill for building test data with [Part Factory](https://partfactory.dev/)
|
|
4
|
+
(`@kensio/part-factory`), a small typed object factory library.
|
|
5
|
+
|
|
6
|
+
The package README is the authority on the API. This skill covers which factory to reach for, when a
|
|
7
|
+
mapped factory earns its place, and where factories should live.
|
|
8
|
+
|
|
9
|
+
## Install
|
|
10
|
+
|
|
11
|
+
From the marketplace:
|
|
12
|
+
|
|
13
|
+
```bash
|
|
14
|
+
claude plugin marketplace add KensioSoftware/kensio.ai
|
|
15
|
+
claude plugin install part-factory-test-data@kensio
|
|
16
|
+
```
|
|
17
|
+
|
|
18
|
+
From npm:
|
|
19
|
+
|
|
20
|
+
```bash
|
|
21
|
+
npm install @kensio/part-factory-test-data
|
|
22
|
+
```
|
|
23
|
+
|
|
24
|
+
## What it covers
|
|
25
|
+
|
|
26
|
+
**Which factory to reach for.** `StaticFactory` for fixed defaults. `DynamicFactory` when the
|
|
27
|
+
defaults have to be generated fresh, paired with [faker](https://fakerjs.dev/), which is what gives
|
|
28
|
+
tests their isolation. `VariantFactory` for a named variation of a base factory, when the variation
|
|
29
|
+
is worth a name in the test. `MappedFactory` when the output shape differs from the parts you want
|
|
30
|
+
to override.
|
|
31
|
+
|
|
32
|
+
**When a mapped factory earns its place.** The map should be a real transformation: parts to an
|
|
33
|
+
encoded form body, or front matter parts to a file as written on disk. If the map is copying fields
|
|
34
|
+
across into an object of the same shape, you wanted a `DynamicFactory`.
|
|
35
|
+
|
|
36
|
+
**Do not wrap a factory in a function that applies overrides.** `make(overrides)` already is that
|
|
37
|
+
function, and it does the job better, since its overrides are partial all the way down the nested
|
|
38
|
+
structure where a spread replaces a nested object whole. A wrapper doing more than passing overrides
|
|
39
|
+
through is a signal that the factory is the wrong shape, or that the output type is fighting you.
|
|
40
|
+
|
|
41
|
+
**Factories belong beside the type they construct.** A library defining an event or message shape
|
|
42
|
+
should export a factory for it, so consumers never hand-roll the literal. The example the skill
|
|
43
|
+
works through is an AWS Lambda function URL invocation event in payload format 2.0: around thirty
|
|
44
|
+
lines of which two matter to any test. Hand-writing it in three files gives three copies that drift,
|
|
45
|
+
and it carries a trap, because the path is in the event twice, as `rawPath` and as
|
|
46
|
+
`requestContext.http.path`. A test that sets one and not the other passes against a handler reading
|
|
47
|
+
the field the test set, and fails in production against the same handler reading the other one. A
|
|
48
|
+
`MappedFactory` sets it once.
|
|
49
|
+
|
|
50
|
+
## Related skills
|
|
51
|
+
|
|
52
|
+
- [`isolated-testing-style`](https://github.com/KensioSoftware/kensio.ai/tree/main/plugins/isolated-testing-style)
|
|
53
|
+
is the general argument for keeping setup out of test bodies, and for taking isolation from
|
|
54
|
+
randomised data.
|
|
55
|
+
- [`yulin-aws-simulation`](https://github.com/KensioSoftware/kensio.ai/tree/main/plugins/yulin-aws-simulation)
|
|
56
|
+
covers the AWS simulator those tests run against.
|
|
57
|
+
|
|
58
|
+
Part of [kensio.ai](https://github.com/KensioSoftware/kensio.ai). Licensed under the Apache License
|
|
59
|
+
2.0. See the [LICENSE](https://github.com/KensioSoftware/kensio.ai/blob/main/LICENSE) in the
|
|
60
|
+
repository root.
|
package/package.json
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
1
|
+
{
|
|
2
|
+
"name": "@kensio/part-factory-test-data",
|
|
3
|
+
"version": "1.3.0",
|
|
4
|
+
"description": "How to build test data with @kensio/part-factory: keeping in the factory everything a test does not care about, passing dependencies at call time so factories stay independent and shareable, and choosing between its static, dynamic, variant and mapped factories.",
|
|
5
|
+
"keywords": [
|
|
6
|
+
"claude",
|
|
7
|
+
"claude-code",
|
|
8
|
+
"claude-code-plugin",
|
|
9
|
+
"factory",
|
|
10
|
+
"faker",
|
|
11
|
+
"kensio",
|
|
12
|
+
"part-factory",
|
|
13
|
+
"skill",
|
|
14
|
+
"test-data",
|
|
15
|
+
"testing"
|
|
16
|
+
],
|
|
17
|
+
"homepage": "https://kensio.ai",
|
|
18
|
+
"license": "Apache-2.0",
|
|
19
|
+
"author": "Kensio Software <hugh@kensiosoftware.co.uk>",
|
|
20
|
+
"repository": {
|
|
21
|
+
"type": "git",
|
|
22
|
+
"url": "git+https://github.com/KensioSoftware/kensio.ai.git",
|
|
23
|
+
"directory": "plugins/part-factory-test-data"
|
|
24
|
+
},
|
|
25
|
+
"files": [
|
|
26
|
+
".claude-plugin",
|
|
27
|
+
"skills",
|
|
28
|
+
"README.md"
|
|
29
|
+
],
|
|
30
|
+
"publishConfig": {
|
|
31
|
+
"access": "public"
|
|
32
|
+
}
|
|
33
|
+
}
|
|
@@ -0,0 +1,235 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: part-factory-test-data
|
|
3
|
+
description: Build test data with @kensio/part-factory, keeping in the factory everything a test does not care about, passing dependencies at call time so factories stay independent, and choosing between StaticFactory, DynamicFactory, VariantFactory, MappedFactory and AsyncMappedFactory. Use when writing test fixtures or builders, when a test file is full of object literals, when a shared event, message or payload shape is being hand-written, when a required field is about to be made optional to ease test setup, and when tempted to wrap a factory in a helper function that applies overrides.
|
|
4
|
+
---
|
|
5
|
+
|
|
6
|
+
# Building test data with Part Factory
|
|
7
|
+
|
|
8
|
+
[Part Factory](https://partfactory.dev/) (`@kensio/part-factory`) builds typed objects for tests: a
|
|
9
|
+
factory holds the defaults, and `make(overrides)` returns an object with the overrides applied down
|
|
10
|
+
through the nested structure.
|
|
11
|
+
|
|
12
|
+
The package README is the authority on the API. This skill covers what to put in a factory, where
|
|
13
|
+
factories should live, and which one to reach for.
|
|
14
|
+
|
|
15
|
+
It serves the `isolated-testing-style` skill. Factories are what make building state inside each
|
|
16
|
+
test cheap enough that nobody reaches for a shared fixture in the first place.
|
|
17
|
+
|
|
18
|
+
## Say only what the test is about
|
|
19
|
+
|
|
20
|
+
A factory defines every value the test does not care about, so the test can state only the values it
|
|
21
|
+
does. That is the whole point of one.
|
|
22
|
+
|
|
23
|
+
Without it, a test opens with twenty lines of construction and it is not clear which of them the
|
|
24
|
+
assertions actually depend on. With it, the lines that are there are the lines that matter:
|
|
25
|
+
|
|
26
|
+
```typescript
|
|
27
|
+
it("charges VAT on the order total", () => {
|
|
28
|
+
// Given an order whose lines add up to a total the assertion depends on.
|
|
29
|
+
const order = orderFactory.make({ lines: [{ price: 1000 }, { price: 2000 }] });
|
|
30
|
+
|
|
31
|
+
// When it is priced.
|
|
32
|
+
const priced = priceOrder(order);
|
|
33
|
+
|
|
34
|
+
// Then VAT is a fifth of that total.
|
|
35
|
+
expect(priced.vat).toBe(600);
|
|
36
|
+
});
|
|
37
|
+
```
|
|
38
|
+
|
|
39
|
+
The prices are written down because the assertion is arithmetic on them. The order id, the customer
|
|
40
|
+
id and the dates are not, because the test would read the same whatever they were. The rule is that
|
|
41
|
+
simple: **if an assertion depends on a value, put it in the test; otherwise let the factory supply
|
|
42
|
+
it.**
|
|
43
|
+
|
|
44
|
+
This also removes a pressure that quietly damages production types. When constructing an object by
|
|
45
|
+
hand is painful, the tempting fix is to mark its required fields optional so tests can skip them —
|
|
46
|
+
weakening the type for every caller in order to serve the tests. A factory makes the setup cheap, so
|
|
47
|
+
the type can go on saying what is actually required.
|
|
48
|
+
|
|
49
|
+
## Overrides are a deep partial
|
|
50
|
+
|
|
51
|
+
Overrides merge into the defaults rather than replacing them, all the way down the nested structure.
|
|
52
|
+
That is what makes "state only what matters" possible: a test can set one field three levels deep
|
|
53
|
+
and leave its siblings alone.
|
|
54
|
+
|
|
55
|
+
Two consequences worth knowing:
|
|
56
|
+
|
|
57
|
+
- Arrays override by index, so `{ lines: [{ price: 1000 }] }` replaces the first default line and
|
|
58
|
+
leaves any others in place.
|
|
59
|
+
- An empty array does not clear the defaults. To assert on emptiness, build the case some other way
|
|
60
|
+
rather than expecting `{ lines: [] }` to do it.
|
|
61
|
+
|
|
62
|
+
## Which factory
|
|
63
|
+
|
|
64
|
+
- **`StaticFactory`** when the defaults are fixed values. The simplest thing that works, and the
|
|
65
|
+
right default choice.
|
|
66
|
+
- **`DynamicFactory`** when the defaults have to be generated fresh for each object. Pair it with
|
|
67
|
+
[`@faker-js/faker`](https://fakerjs.dev/). This is what gives tests their isolation: a random
|
|
68
|
+
email or a UUID means two tests cannot collide, so neither needs tearing down.
|
|
69
|
+
- **`VariantFactory`** for a named variation of a base factory, when the variation is a concept the
|
|
70
|
+
tests talk about. `closedOfferFactory` reads better in ten tests than
|
|
71
|
+
`offerFactory.make({ closesAt: aMinuteAgo })` written ten times.
|
|
72
|
+
- **`MappedFactory`** when the output shape differs from the parts you want to override.
|
|
73
|
+
- **`AsyncMappedFactory`** when producing the value means awaiting something, such as inserting a
|
|
74
|
+
row or signing a payload.
|
|
75
|
+
|
|
76
|
+
```typescript
|
|
77
|
+
import { DynamicFactory } from "@kensio/part-factory";
|
|
78
|
+
import { faker } from "@faker-js/faker";
|
|
79
|
+
|
|
80
|
+
export const customerFactory = new DynamicFactory<Customer>(() => ({
|
|
81
|
+
id: faker.string.uuid(),
|
|
82
|
+
email: faker.internet.email(),
|
|
83
|
+
name: faker.person.fullName(),
|
|
84
|
+
}));
|
|
85
|
+
```
|
|
86
|
+
|
|
87
|
+
## Do not add a variant for every case
|
|
88
|
+
|
|
89
|
+
A variant earns its name when several tests mean the same thing by it. One test that needs an
|
|
90
|
+
unusual value should write that value down, not gain a factory of its own in a file somewhere else.
|
|
91
|
+
|
|
92
|
+
The failure mode is a directory of `cancelledOrderWithRefundAndNoAddressFactory` names, each used
|
|
93
|
+
once, where reading a test means going to find out what its factory actually sets. Writing the field
|
|
94
|
+
in the test is shorter and says more.
|
|
95
|
+
|
|
96
|
+
## Reach for MappedFactory only when the map is a real transformation
|
|
97
|
+
|
|
98
|
+
`MappedFactory` earns its place when the thing you want to override is not shaped like the thing you
|
|
99
|
+
want back. Good cases:
|
|
100
|
+
|
|
101
|
+
- Parts to an encoded form body: `{ email, password }` mapped to a
|
|
102
|
+
`application/x-www-form-urlencoded` string.
|
|
103
|
+
- Front matter parts to a file as written on disk: `{ title, tags, body }` mapped to the YAML block
|
|
104
|
+
plus the markdown underneath.
|
|
105
|
+
- Components to a formatted identifier: ARN parts mapped to the ARN string.
|
|
106
|
+
|
|
107
|
+
If the mapping function is copying fields across into an object of the same shape, you wanted a
|
|
108
|
+
`DynamicFactory`. An identity map adds a type parameter, a second function and a layer of
|
|
109
|
+
indirection, and buys nothing.
|
|
110
|
+
|
|
111
|
+
## Pass dependencies at call time
|
|
112
|
+
|
|
113
|
+
A factory that needs something from the outside world — a store, a client, a configured host —
|
|
114
|
+
declares it as a third type parameter and receives it as the second argument to `make`:
|
|
115
|
+
|
|
116
|
+
```typescript
|
|
117
|
+
export const storedOrderFactory = new AsyncMappedFactory<
|
|
118
|
+
OrderParts,
|
|
119
|
+
Order,
|
|
120
|
+
{ orders: OrderStore }
|
|
121
|
+
>(
|
|
122
|
+
() => ({ total: faker.number.int({ min: 100, max: 10_000 }) }),
|
|
123
|
+
async (parts, { orders }) => orders.insert({ id: faker.string.uuid(), ...parts }),
|
|
124
|
+
);
|
|
125
|
+
|
|
126
|
+
// In a test, which decides what to hand it.
|
|
127
|
+
const order = await storedOrderFactory.make({ total: 5000 }, { orders });
|
|
128
|
+
```
|
|
129
|
+
|
|
130
|
+
Dependencies are given at call time rather than held by the factory, and are used as they are given,
|
|
131
|
+
never fetched or awaited. That is what keeps a factory shareable: it holds no state of its own,
|
|
132
|
+
reaches for nothing ambient, and cannot depend on what another factory did first. A factory built
|
|
133
|
+
this way can be used in every test file in the codebase and still stand on its own.
|
|
134
|
+
|
|
135
|
+
Keep each dependency as narrow as the factory actually needs. An `OrderStore` is a dependency; the
|
|
136
|
+
whole application is not. A wide dependency is how state starts leaking between tests that were
|
|
137
|
+
supposed to be independent.
|
|
138
|
+
|
|
139
|
+
## Do not wrap a factory in a function that applies overrides
|
|
140
|
+
|
|
141
|
+
```typescript
|
|
142
|
+
// Anti-pattern. This is make(overrides) with extra steps.
|
|
143
|
+
export function makeCustomer(overrides: Partial<Customer> = {}): Customer {
|
|
144
|
+
return { ...customerFactory.make(), ...overrides };
|
|
145
|
+
}
|
|
146
|
+
```
|
|
147
|
+
|
|
148
|
+
`make(overrides)` already is that function, and it does the job better: its overrides are partial
|
|
149
|
+
all the way down the nested structure, where the spread above replaces a nested object whole.
|
|
150
|
+
|
|
151
|
+
A wrapper that does anything more than pass overrides through is a signal to read rather than to
|
|
152
|
+
write. It usually means one of two things:
|
|
153
|
+
|
|
154
|
+
- **The factory is the wrong shape.** The wrapper is computing something the defaults should be
|
|
155
|
+
computing, or deriving one field from another. Move that into a `DynamicFactory` defaults
|
|
156
|
+
function, which receives the overrides, or into a `MappedFactory` map.
|
|
157
|
+
- **The output type is fighting you.** The wrapper is casting, widening or filling in a field the
|
|
158
|
+
type demands but the test does not care about. Fix the type, or use `MappedFactory` so the parts
|
|
159
|
+
and the output are allowed to differ.
|
|
160
|
+
|
|
161
|
+
The same goes for a wrapper that exists to pass a dependency: declare it on the factory instead.
|
|
162
|
+
|
|
163
|
+
## Factories belong beside the type they construct
|
|
164
|
+
|
|
165
|
+
A library that defines an event, a message or a payload shape should export a factory for it.
|
|
166
|
+
Otherwise every consumer hand-rolls the literal, and every copy drifts.
|
|
167
|
+
|
|
168
|
+
The worked example: an AWS Lambda function URL invocation event, payload format 2.0. It is around
|
|
169
|
+
thirty lines, of which two matter to any given test.
|
|
170
|
+
|
|
171
|
+
```json
|
|
172
|
+
{
|
|
173
|
+
"version": "2.0",
|
|
174
|
+
"routeKey": "$default",
|
|
175
|
+
"rawPath": "/upload",
|
|
176
|
+
"rawQueryString": "part=3",
|
|
177
|
+
"headers": { "host": "abc.lambda-url.eu-west-2.on.aws", "user-agent": "..." },
|
|
178
|
+
"queryStringParameters": { "part": "3" },
|
|
179
|
+
"cookies": ["session=abc"],
|
|
180
|
+
"requestContext": {
|
|
181
|
+
"http": { "method": "POST", "path": "/upload", "sourceIp": "127.0.0.1" }
|
|
182
|
+
},
|
|
183
|
+
"body": "...",
|
|
184
|
+
"isBase64Encoded": false
|
|
185
|
+
}
|
|
186
|
+
```
|
|
187
|
+
|
|
188
|
+
Hand-writing that in three test files gives three copies that drift as the shape changes. It also
|
|
189
|
+
carries a real trap: the path is in the event twice, as `rawPath` and as `requestContext.http.path`,
|
|
190
|
+
and the query string is in it twice, as `rawQueryString` and as `queryStringParameters`. A test that
|
|
191
|
+
sets one and not the other passes against a handler reading the field the test set, and fails in
|
|
192
|
+
production against the same handler reading the other one.
|
|
193
|
+
|
|
194
|
+
A `MappedFactory` removes both problems. The parts are what a test cares about, and the map is what
|
|
195
|
+
fills the event in consistently:
|
|
196
|
+
|
|
197
|
+
```typescript
|
|
198
|
+
import { MappedFactory } from "@kensio/part-factory";
|
|
199
|
+
|
|
200
|
+
interface FunctionUrlRequestParts {
|
|
201
|
+
method: string;
|
|
202
|
+
path: string;
|
|
203
|
+
query: Record<string, string>;
|
|
204
|
+
body?: string;
|
|
205
|
+
}
|
|
206
|
+
|
|
207
|
+
export const functionUrlEventFactory = new MappedFactory<
|
|
208
|
+
FunctionUrlRequestParts,
|
|
209
|
+
FunctionUrlEvent
|
|
210
|
+
>(
|
|
211
|
+
() => ({ method: "GET", path: "/", query: {} }),
|
|
212
|
+
(parts) => ({
|
|
213
|
+
version: "2.0",
|
|
214
|
+
routeKey: "$default",
|
|
215
|
+
rawPath: parts.path,
|
|
216
|
+
rawQueryString: new URLSearchParams(parts.query).toString(),
|
|
217
|
+
queryStringParameters: parts.query,
|
|
218
|
+
// ... and the rest, filled in once.
|
|
219
|
+
requestContext: {
|
|
220
|
+
http: { method: parts.method, path: parts.path, sourceIp: "127.0.0.1" },
|
|
221
|
+
},
|
|
222
|
+
body: parts.body,
|
|
223
|
+
isBase64Encoded: false,
|
|
224
|
+
}),
|
|
225
|
+
);
|
|
226
|
+
|
|
227
|
+
// In a test, the two lines that matter are the two lines written.
|
|
228
|
+
const event = functionUrlEventFactory.make({ method: "POST", path: "/upload" });
|
|
229
|
+
```
|
|
230
|
+
|
|
231
|
+
The path can no longer disagree with itself, because there is one place it is set.
|
|
232
|
+
|
|
233
|
+
Export the factory from the package that owns the type, from a test-support entry point so it does
|
|
234
|
+
not ship in the production bundle. Consumers then get a correct event with one call, and a change to
|
|
235
|
+
the shape is made once.
|