@functional-systems/lambdadb-cli 0.1.4-dev.16 → 0.1.4-dev.18
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/CHANGELOG.md +7 -2
- package/CONTRIBUTING.md +95 -10
- package/DESIGN.md +11 -3
- package/README.md +14 -11
- package/RELEASING.md +17 -169
- package/package.json +3 -4
- package/VALIDATION.md +0 -1017
package/CHANGELOG.md
CHANGED
|
@@ -4,9 +4,15 @@
|
|
|
4
4
|
|
|
5
5
|
### Maintenance
|
|
6
6
|
|
|
7
|
+
- Standardize feature descriptions to native embedding and native reranking;
|
|
8
|
+
preserve legacy wire fields, CLI behavior and historical references.
|
|
7
9
|
- Synchronize the 0.1.3 release history and start the 0.1.4 development line.
|
|
8
10
|
- Prepare the Homebrew formula for verified stable 0.1.3 and record publication evidence.
|
|
9
11
|
|
|
12
|
+
### Dependencies
|
|
13
|
+
|
|
14
|
+
- Pin `@functional-systems/lambdadb` and the npm lockfile to stable 0.8.1.
|
|
15
|
+
|
|
10
16
|
## [0.1.3] - 2026-10-06
|
|
11
17
|
|
|
12
18
|
### Added
|
|
@@ -148,6 +154,5 @@ Published publicly on npm's `dev` channel on 2026-09-18 as the manual bootstrap.
|
|
|
148
154
|
The initial OIDC build `0.1.0-dev.4` followed from the same reviewed develop
|
|
149
155
|
commit; generated dev versions do not create separate source changelog entries.
|
|
150
156
|
Automatic dev publication is enabled. The observed `latest` tag remains at the
|
|
151
|
-
bootstrap until a reviewed stable release replaces it.
|
|
152
|
-
installation and development-project smoke evidence are recorded in VALIDATION.md.
|
|
157
|
+
bootstrap until a reviewed stable release replaces it.
|
|
153
158
|
Main promotion is required only for rc/stable.
|
package/CONTRIBUTING.md
CHANGED
|
@@ -66,16 +66,12 @@ development base in that synchronization PR so develop retains a dev version.
|
|
|
66
66
|
|
|
67
67
|
A separate `release/*` branch can be created from reviewed develop and target
|
|
68
68
|
main when preparing rc/stable metadata while develop retains its automatic dev
|
|
69
|
-
base.
|
|
70
|
-
|
|
71
|
-
|
|
69
|
+
base. Do not merge stable package metadata into develop without setting its next
|
|
70
|
+
dev base, because automatic development publication requires an `X.Y.Z-dev.N`
|
|
71
|
+
version.
|
|
72
72
|
|
|
73
73
|
The first npm bootstrap, rc/stable tags and GitHub Releases remain explicit
|
|
74
|
-
release actions. Subsequent dev publication is automatic once enabled.
|
|
75
|
-
Bootstrap `0.1.0-dev.1` and the first OIDC development publication `0.1.0-dev.4`
|
|
76
|
-
were verified on 2026-09-18; automatic dev publication is enabled. Publication,
|
|
77
|
-
provenance and development-project smoke evidence are recorded in
|
|
78
|
-
[VALIDATION.md](VALIDATION.md). Follow
|
|
74
|
+
release actions. Subsequent dev publication is automatic once enabled. Follow
|
|
79
75
|
[RELEASING.md](RELEASING.md) for package preparation, version/channel rules,
|
|
80
76
|
credentialed live evidence, initial npm bootstrap and Trusted Publishing.
|
|
81
77
|
Local tests and PR validation alone never trigger publication.
|
|
@@ -89,8 +85,8 @@ review requirements.
|
|
|
89
85
|
Never commit project keys, `.env` files or signed URLs. Runtime tests use loopback
|
|
90
86
|
servers with synthetic credentials. Live tests require an explicitly designated
|
|
91
87
|
development project and credential set; local or CI success is not proof of
|
|
92
|
-
live-service readiness. See [README.md](README.md) and
|
|
93
|
-
[
|
|
88
|
+
live-service readiness. See [README.md](README.md) for public behavior and the
|
|
89
|
+
[validation scope](#validation-scope) below for evidence boundaries.
|
|
94
90
|
|
|
95
91
|
## Compatibility and maintenance
|
|
96
92
|
|
|
@@ -104,3 +100,92 @@ Review dependency updates through the same PR checks, including the actual SDK
|
|
|
104
100
|
contract and installed-package tests. Keep workflow actions pinned to commit
|
|
105
101
|
SHAs and document version updates in their comments. Do not copy another
|
|
106
102
|
repository's runtime matrix or release tooling without checking this CLI's needs.
|
|
103
|
+
|
|
104
|
+
## Validation scope
|
|
105
|
+
|
|
106
|
+
- `npm test` builds the CLI and executes unit, subprocess and release-tooling
|
|
107
|
+
contracts. Loopback HTTP servers exercise the installed SDK without contacting
|
|
108
|
+
LambdaDB: input and target precedence, query and Collection mapping, signed
|
|
109
|
+
transfers, pagination, complete results, credential redaction and exit codes.
|
|
110
|
+
Failure cases cover bounded inputs, timeouts, cancellation, partial and unknown
|
|
111
|
+
writes, and the absence of mutation retries.
|
|
112
|
+
- `npm run test:package` packs the candidate, checks its inventory and executable,
|
|
113
|
+
installs it into a temporary consumer and repeats the CLI contracts against
|
|
114
|
+
that artifact. To check an existing candidate, use
|
|
115
|
+
`npm run test:package -- /path/to/package.tgz`. Building test sources alone is
|
|
116
|
+
not evidence that tests executed.
|
|
117
|
+
- [Homebrew validation](packaging/homebrew/README.md#local-and-ci-verification)
|
|
118
|
+
checks the stable formula's installation and released CLI contracts on macOS
|
|
119
|
+
and Linux. A new installation does not establish an existing-user upgrade;
|
|
120
|
+
verify upgrades separately as part of the tap handoff.
|
|
121
|
+
- Local tests and PR CI do not establish backend deployment provenance, live
|
|
122
|
+
search quality, production performance, billing readiness, exactly-once writes
|
|
123
|
+
or durable resume. Check the deployed backend revision separately when a
|
|
124
|
+
feature requires server support. The CI matrix does not validate Windows
|
|
125
|
+
permissions or signals.
|
|
126
|
+
- Input byte/document limits bound accepted input, not process memory. Large
|
|
127
|
+
mock responses and a bounded live sample are not exhaustive memory or storage
|
|
128
|
+
policy tests. A successful live sample applies only to its designated target,
|
|
129
|
+
revision, refs and inputs.
|
|
130
|
+
|
|
131
|
+
## Live validation
|
|
132
|
+
|
|
133
|
+
Provision a disposable development project/key outside this tool, with Collection
|
|
134
|
+
create/update/delete, ordinary and bulk write, and query/fetch access.
|
|
135
|
+
Supply the key through secure shell input or a secret manager, never in history.
|
|
136
|
+
|
|
137
|
+
```sh
|
|
138
|
+
export LAMBDADB_ENDPOINT=https://YOUR_DEV_API_ORIGIN
|
|
139
|
+
export LAMBDADB_PROJECT=YOUR_DEV_PROJECT
|
|
140
|
+
# Inject LAMBDADB_API_KEY securely; no .env file is loaded automatically.
|
|
141
|
+
export LAMBDADB_RUN_LIVE_TESTS=1
|
|
142
|
+
export LAMBDADB_LIVE_CONFIRM_PROJECT="$LAMBDADB_PROJECT"
|
|
143
|
+
npm run test:live
|
|
144
|
+
```
|
|
145
|
+
|
|
146
|
+
The test fails before API calls when required settings are absent. It creates a
|
|
147
|
+
random temporary collection, imports two 3 MiB documents plus a bulk document,
|
|
148
|
+
accepts and verifies metadata for all 49 fixed analyzer presets, checks exact
|
|
149
|
+
facet-only and document+facet buckets, and checks committed query/fetch contents
|
|
150
|
+
with a maximum 300-second observation window per stage. It also checks native
|
|
151
|
+
reranking with default criteria, null and custom criteria, including
|
|
152
|
+
final/retrieval scores and status metadata.
|
|
153
|
+
It creates two additional temporary Collections for native embedding-only and
|
|
154
|
+
legacy true inputs. It exercises actual CLI create/update, normalized metadata,
|
|
155
|
+
document/query embedding generation, ordinary retrieval,
|
|
156
|
+
Bayesian candidate budgets, default/explicit rerank budgets, null rerank and
|
|
157
|
+
server-side rejection of invalid Bayesian requests. All three Collections are
|
|
158
|
+
cleaned up with subsequent HTTP 404 verification.
|
|
159
|
+
|
|
160
|
+
Request timeouts and polling sleeps are capped to the remaining budget, and
|
|
161
|
+
responses at or after the deadline cannot pass. Process termination
|
|
162
|
+
and event-loop scheduling may delay reporting. It reports elapsed time and
|
|
163
|
+
matched-document counts; a timeout is a failed smoke, not proof that accepted
|
|
164
|
+
writes were rejected or lost.
|
|
165
|
+
|
|
166
|
+
Collection cleanup uses the SDK because deletion is outside the public CLI MVP.
|
|
167
|
+
Only the test's own Collections are eligible for cleanup. An interrupted process
|
|
168
|
+
or failed cleanup may leave those Collections; inspect them manually.
|
|
169
|
+
|
|
170
|
+
This sample does not prove global index readiness, ranking quality, every ref's
|
|
171
|
+
live behavior, or exactly-once writes. A large result alone does not prove the
|
|
172
|
+
service selected docsUrl; mock tests explicitly exercise that transport path.
|
|
173
|
+
Use existing known development tag/alias refs for additional live ref evidence.
|
|
174
|
+
|
|
175
|
+
## Recording verification results
|
|
176
|
+
|
|
177
|
+
Keep repeatable procedures and supported behavior in repository documentation.
|
|
178
|
+
Put each change's exact commands, outcomes, source/artifact revision, CI links,
|
|
179
|
+
live target and cleanup result, and material unperformed checks in its PR.
|
|
180
|
+
For publication, include the verified version, integrity and provenance summary
|
|
181
|
+
in the release PR or GitHub Release. User-facing changes and compatibility notes
|
|
182
|
+
belong in the changelog and release notes.
|
|
183
|
+
|
|
184
|
+
Retain sanitized evidence needed for long-term review as release assets or in
|
|
185
|
+
an approved durable location; CI logs and artifacts have retention limits, and
|
|
186
|
+
local temporary paths are not durable references. Never include credentials or
|
|
187
|
+
signed URLs. Historical repository records remain available through Git history.
|
|
188
|
+
|
|
189
|
+
Do not append per-run logs, release-status snapshots or successful test counts to
|
|
190
|
+
maintainer guides. A completed verification does not require another documentation
|
|
191
|
+
PR. Update these guides when procedures, requirements or supported behavior change.
|
package/DESIGN.md
CHANGED
|
@@ -56,8 +56,9 @@ preserves `facets` while removing the consumed signed URL. Facet field names joi
|
|
|
56
56
|
other user-defined maps in credential redaction, including collision preservation;
|
|
57
57
|
fixed facet result/bucket keys retain their SDK schema even for short credentials.
|
|
58
58
|
Loopback subprocess tests cover serialization and output through the published SDK,
|
|
59
|
-
and the installed-tarball suite repeats these contracts. Live
|
|
60
|
-
|
|
59
|
+
and the installed-tarball suite repeats these contracts. Live validation is
|
|
60
|
+
separate; follow the [contributor procedure](CONTRIBUTING.md#live-validation)
|
|
61
|
+
and record results in the relevant PR.
|
|
61
62
|
|
|
62
63
|
## SDK 0.7.0 contract update (2026-10-04)
|
|
63
64
|
|
|
@@ -128,6 +129,13 @@ fetch/query already delegate docsUrl to the SDK and preserve full results. Impor
|
|
|
128
129
|
retains its existing bounded preflight and ordinary/bulk behavior; native
|
|
129
130
|
embeddings use ordinary upsert. No changes to those flows were needed.
|
|
130
131
|
|
|
132
|
+
## SDK 0.8.1 terminology update
|
|
133
|
+
|
|
134
|
+
The runtime and lockfile now pin `@functional-systems/lambdadb@0.8.1`.
|
|
135
|
+
Feature descriptions use native embedding and native reranking. The legacy
|
|
136
|
+
`managedEmbedding` wire field and existing CLI behavior remain unchanged.
|
|
137
|
+
The versioned contracts above retain their original source paths.
|
|
138
|
+
|
|
131
139
|
## Architecture
|
|
132
140
|
|
|
133
141
|
```text
|
|
@@ -170,7 +178,7 @@ helpers own presigned response downloads and authentication isolation.
|
|
|
170
178
|
| No readiness wait or version-management commands | First use works through main; reads can target existing refs, without assuming lifecycle or indexing guarantees. |
|
|
171
179
|
|
|
172
180
|
The public API documents ordinary upsert at 6 MB and bulk at 200 MB; CLI limits
|
|
173
|
-
are deliberately smaller. Bulk is unsupported for
|
|
181
|
+
are deliberately smaller. Bulk is unsupported for native embedding vector
|
|
174
182
|
fields. `consistentRead` only supports direct branches and excludes pending bulk
|
|
175
183
|
imports. Collection statistics describe committed main, not selected ref readiness.
|
|
176
184
|
|
package/README.md
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
# LambdaDB CLI
|
|
2
2
|
|
|
3
3
|
A first, project-scoped CLI for developers, coding agents and CI. It uses
|
|
4
|
-
`@functional-systems/lambdadb@0.8.
|
|
4
|
+
`@functional-systems/lambdadb@0.8.1` for authentication, HTTP, read retries,
|
|
5
5
|
pagination, bulk transfers and large response downloads.
|
|
6
6
|
|
|
7
7
|
## Install and run
|
|
@@ -22,8 +22,8 @@ For reproducible CI runs, pin an exact published version such as
|
|
|
22
22
|
`@functional-systems/lambdadb-cli@0.1.3`. To try development builds, explicitly use
|
|
23
23
|
`@functional-systems/lambdadb-cli@dev`; this moving channel contains prereleases.
|
|
24
24
|
Installed CLIs do not update themselves. See
|
|
25
|
-
[
|
|
26
|
-
[
|
|
25
|
+
[GitHub Releases](https://github.com/lambdadb/lambdadb-cli/releases) and
|
|
26
|
+
[CHANGELOG.md](CHANGELOG.md) for published changes and compatibility notes.
|
|
27
27
|
|
|
28
28
|
### Homebrew
|
|
29
29
|
|
|
@@ -235,7 +235,7 @@ backend `9072a1bc8925954369a887f558f1eaf387b7ea0e`. See the SDK 0.8.0
|
|
|
235
235
|
[Bayesian contract](https://github.com/lambdadb/lambdadb-typescript-client/blob/v0.8.0/docs/bayesian-search.md)
|
|
236
236
|
and [native embedding contract](https://github.com/lambdadb/lambdadb-typescript-client/blob/v0.8.0/docs/native-embeddings.md).
|
|
237
237
|
|
|
238
|
-
###
|
|
238
|
+
### Native reranking
|
|
239
239
|
|
|
240
240
|
```sh
|
|
241
241
|
lambdadb query --collection cli-demo-docs --ref branch:main \
|
|
@@ -243,7 +243,8 @@ lambdadb query --collection cli-demo-docs --ref branch:main \
|
|
|
243
243
|
```
|
|
244
244
|
|
|
245
245
|
Add optional `rerank` to a query file; omission/null keeps existing searches.
|
|
246
|
-
This is a per-query setting
|
|
246
|
+
This is a per-query setting. The LambdaDB server supplies provider credentials;
|
|
247
|
+
you do not supply a Jev key.
|
|
247
248
|
Use `provider: "typesafe"`, `model: "jev-1.13.0"`, nonblank `queryText` (also for
|
|
248
249
|
raw vectors, up to 8 KiB UTF-8), and 1–8 unique stored scalar text `fields`.
|
|
249
250
|
Returned-field projection remains independent of these model inputs.
|
|
@@ -306,7 +307,7 @@ and 16,000,000 bytes; `--batch-bytes` allows up to 64,000,000. These are conserv
|
|
|
306
307
|
CLI limits, not advertised server maxima. The SDK also enforces the bulk limit
|
|
307
308
|
returned by the server. `--mode bulk` calls the existing SDK `bulkUpsertDocs`
|
|
308
309
|
helper with the selected branch, including signed upload headers and finalization.
|
|
309
|
-
Bulk is unsupported for
|
|
310
|
+
Bulk is unsupported for native embedding vector fields; use ordinary upsert.
|
|
310
311
|
|
|
311
312
|
Import result fields are stable under `schemaVersion: 1`:
|
|
312
313
|
|
|
@@ -383,7 +384,7 @@ this SDK update. If migration is needed, deliberately create a new Collection an
|
|
|
383
384
|
reinsert the source data; the CLI performs no automatic migration. Accepted imports
|
|
384
385
|
and `consistentRead` are not proof of committed facet/index readiness.
|
|
385
386
|
|
|
386
|
-
This source pins SDK 0.8.
|
|
387
|
+
This source pins SDK 0.8.1. Existing CLI installations keep their packaged SDK
|
|
387
388
|
until a new CLI version is published and installed.
|
|
388
389
|
|
|
389
390
|
## Output, deadlines and exit codes
|
|
@@ -447,7 +448,8 @@ Tests execute the real CLI and installed SDK against loopback HTTP servers. They
|
|
|
447
448
|
cover the complete first-use flow, ref forwarding, signed transfers, pagination,
|
|
448
449
|
partial and unknown writes, validation, output, credentials and exit codes.
|
|
449
450
|
They do not contact a LambdaDB service. See [DESIGN.md](DESIGN.md) for contract
|
|
450
|
-
|
|
451
|
+
decisions and [CONTRIBUTING.md](CONTRIBUTING.md#validation-scope) for validation
|
|
452
|
+
methods and their limits.
|
|
451
453
|
|
|
452
454
|
This MVP intentionally omits Branch/Tag/Alias management: default `main` supports
|
|
453
455
|
first use, while reads can select existing refs. It also omits Console management,
|
|
@@ -457,9 +459,10 @@ plugins and publication. There are no changes to the SDK or reference repositori
|
|
|
457
459
|
Live verification requires an explicitly designated development project, endpoint
|
|
458
460
|
and project key. Do not discover or borrow credentials from another repository.
|
|
459
461
|
Use `npm run test:live` with the explicit opt-in settings in
|
|
460
|
-
[
|
|
461
|
-
|
|
462
|
-
the SDK to
|
|
462
|
+
[CONTRIBUTING.md](CONTRIBUTING.md#live-validation). It exercises ordinary/bulk
|
|
463
|
+
import, query/fetch, native embeddings and Bayesian reranking in temporary
|
|
464
|
+
Collections, then uses the SDK to delete them and verify their absence. The normal
|
|
465
|
+
test suite and CI never run it.
|
|
463
466
|
|
|
464
467
|
## License
|
|
465
468
|
|
package/RELEASING.md
CHANGED
|
@@ -1,111 +1,9 @@
|
|
|
1
1
|
# Releasing LambdaDB CLI
|
|
2
2
|
|
|
3
|
-
|
|
4
|
-
|
|
5
|
-
|
|
6
|
-
|
|
7
|
-
[GitHub Release v0.1.3](https://github.com/lambdadb/lambdadb-cli/releases/tag/v0.1.3)
|
|
8
|
-
and the [successful OIDC workflow](https://github.com/lambdadb/lambdadb-cli/actions/runs/37466450036).
|
|
9
|
-
It pins SDK 0.8.0 and adds Bayesian queries, native embedding configuration and
|
|
10
|
-
Collection index updates, preserving existing query defaults and error/output contracts.
|
|
11
|
-
|
|
12
|
-
An independent unqualified npm installation selected 0.1.3 and SDK 0.8.0, passed
|
|
13
|
-
all 54 CLI contracts and verified integrity, four registry signatures and three
|
|
14
|
-
attestations. Decoded provenance identifies the release commit, tag and workflow.
|
|
15
|
-
The published tarball is byte-identical to the reviewed candidate, whose installed
|
|
16
|
-
runtime passed development-service validation. At verification, latest=0.1.3 and
|
|
17
|
-
dev=0.1.3-dev.15. See [publication evidence](VALIDATION.md#cli-013-publication-2026-10-06).
|
|
18
|
-
|
|
19
|
-
The post-release synchronization starts the `0.1.4-dev.1` development base and
|
|
20
|
-
prepares the 0.1.3 Homebrew formula. The public tap remains at 0.1.2 until its
|
|
21
|
-
separate PR is reviewed and merged; CLI publication does not update that tap.
|
|
22
|
-
|
|
23
|
-
## CLI 0.1.2 publication
|
|
24
|
-
|
|
25
|
-
CLI 0.1.2 was published on 2026-10-04 from merged main commit
|
|
26
|
-
`6628fe0ed1f405c88fbd91d591e5223d3250dd73` through
|
|
27
|
-
[GitHub Release v0.1.2](https://github.com/lambdadb/lambdadb-cli/releases/tag/v0.1.2)
|
|
28
|
-
and the [successful OIDC workflow](https://github.com/lambdadb/lambdadb-cli/actions/runs/37191188846).
|
|
29
|
-
It pins SDK 0.7.0 and includes 49 fixed analyzer presets, optional per-query
|
|
30
|
-
managed reranking and credential-safe fixed rerank metadata.
|
|
31
|
-
|
|
32
|
-
At stable publication, independent registry verification confirmed latest=0.1.2
|
|
33
|
-
and dev=0.1.2-dev.12.
|
|
34
|
-
The published tarball is byte-identical to the reviewed, live-tested candidate.
|
|
35
|
-
A clean registry consumer passed all 46 CLI contracts, version/help, SDK version,
|
|
36
|
-
integrity, four registry signatures and three attestations. Decoded provenance
|
|
37
|
-
identifies the release commit, tag and workflow. See
|
|
38
|
-
[publication evidence](VALIDATION.md#cli-012-publication-2026-10-04).
|
|
39
|
-
|
|
40
|
-
[PR #16](https://github.com/lambdadb/lambdadb-cli/pull/16) synchronized main into
|
|
41
|
-
develop with the 0.1.3-dev.1 base. Its successful automatic publication produced
|
|
42
|
-
0.1.3-dev.13; an independent registry consumer passed all 46 CLI contracts,
|
|
43
|
-
integrity, signature and provenance checks. At verification, dev=0.1.3-dev.13
|
|
44
|
-
and latest=0.1.2.
|
|
45
|
-
|
|
46
|
-
[Homebrew PR #6](https://github.com/lambdadb/homebrew-tap/pull/6) merged the matching
|
|
47
|
-
0.1.2 formula. Public-tap installation passed on macOS arm64 and Linux x86_64.
|
|
48
|
-
A separate disposable-tap macOS arm64 upgrade from 0.1.1 to 0.1.2 also passed;
|
|
49
|
-
existing-user public-tap upgrades and Linux upgrades were not tested. See
|
|
50
|
-
[post-release evidence](VALIDATION.md#cli-012-post-release-verification-2026-10-04).
|
|
51
|
-
Publication and bounded development smoke do not establish production deployment,
|
|
52
|
-
search quality, load/failure coverage or billing readiness.
|
|
53
|
-
|
|
54
|
-
## CLI 0.1.1 publication
|
|
55
|
-
|
|
56
|
-
CLI 0.1.1 was published on 2026-09-29 through
|
|
57
|
-
[GitHub Release v0.1.1](https://github.com/lambdadb/lambdadb-cli/releases/tag/v0.1.1)
|
|
58
|
-
and the [OIDC release workflow](https://github.com/lambdadb/lambdadb-cli/actions/runs/36557578359).
|
|
59
|
-
Tag/main commit `80a4c10628ad2ad4ebf8d90a01924d27dcb4a790` has the same tree as
|
|
60
|
-
reviewed candidate `0291970`. The workflow passed on attempt 1, including Node
|
|
61
|
-
22/24 validation and tests of the exact publication tarball.
|
|
62
|
-
|
|
63
|
-
Registry reads confirmed `latest=0.1.1`, while `dev=0.1.1-dev.10` remained unchanged.
|
|
64
|
-
A clean unqualified npm installation selected CLI 0.1.1 and SDK 0.6.0, passed all
|
|
65
|
-
40 CLI contracts and verified registry signatures and attestations. The published
|
|
66
|
-
tarball is byte-identical to the tested final candidate; provenance identifies the
|
|
67
|
-
release commit, tag and workflow. See [VALIDATION.md](VALIDATION.md#cli-011-publication).
|
|
68
|
-
|
|
69
|
-
The 0.1.1 post-release synchronization started the `0.1.2-dev.1` development base
|
|
70
|
-
and prepared the 0.1.1 Homebrew formula. The public tap matched that formula at
|
|
71
|
-
the pre-release check on 2026-10-04; the completed 0.1.2 handoff is recorded above.
|
|
72
|
-
Facets still require a supporting server and newly built keyword indexes;
|
|
73
|
-
publication does not migrate Collections or upgrade installed CLIs.
|
|
74
|
-
|
|
75
|
-
## Prior publication status
|
|
76
|
-
|
|
77
|
-
The first stable version, `0.1.0`, was published on 2026-09-19 through
|
|
78
|
-
[GitHub Release v0.1.0](https://github.com/lambdadb/lambdadb-cli/releases/tag/v0.1.0)
|
|
79
|
-
and the [OIDC release workflow](https://github.com/lambdadb/lambdadb-cli/actions/runs/35421226854).
|
|
80
|
-
Tag and main release commit `c8d389f` have the same tree as the reviewed and
|
|
81
|
-
live-tested candidate `658ea7c`. The workflow passed on attempt 1, including Node
|
|
82
|
-
22/24 validation and tests of the exact publication tarball.
|
|
83
|
-
|
|
84
|
-
Registry verification confirmed `latest=0.1.0` and `dev=0.1.0-dev.5`. A clean,
|
|
85
|
-
unqualified npm install passed all 35 CLI contracts, version/help, integrity,
|
|
86
|
-
provenance and signature checks. Version/tag and attestation metadata propagated
|
|
87
|
-
separately; verification used read retries without republishing.
|
|
88
|
-
|
|
89
|
-
The post-release synchronization merged main history back into develop with
|
|
90
|
-
`0.1.1-dev.1` as its development base. After the Homebrew documentation merge,
|
|
91
|
-
[develop CI](https://github.com/lambdadb/lambdadb-cli/actions/runs/35434313586)
|
|
92
|
-
published and verified `0.1.1-dev.8` from `225dec3` on 2026-09-19. Registry and
|
|
93
|
-
provenance metadata identify that commit; `latest` remains `0.1.0`. Development
|
|
94
|
-
builds do not require a separate stable release or tag.
|
|
95
|
-
|
|
96
|
-
Bootstrap `0.1.0-dev.1` and the first OIDC development builds were verified on
|
|
97
|
-
2026-09-18. The ordinary develop merge published `0.1.0-dev.5` in
|
|
98
|
-
[one successful attempt](https://github.com/lambdadb/lambdadb-cli/actions/runs/35344153978),
|
|
99
|
-
handling approximately three minutes of registry propagation without another write.
|
|
100
|
-
The bootstrap initially assigned `latest` to `0.1.0-dev.1`, and a removal attempt
|
|
101
|
-
returned HTTP 400. The stable publication replaces that temporary tag state.
|
|
102
|
-
Do not repeat bootstrap or change tags to work around propagation delays.
|
|
103
|
-
|
|
104
|
-
These are dated validation milestones, not a continuously updated version list.
|
|
105
|
-
Check npm for current dist-tags. Apache-2.0 licensing, publication, provenance,
|
|
106
|
-
consumer checks and development-project smoke evidence are recorded in
|
|
107
|
-
[VALIDATION.md](VALIDATION.md). Automatic dev publication is enabled;
|
|
108
|
-
future stable/rc tags and GitHub Releases remain explicit release actions.
|
|
3
|
+
This guide defines release policy and procedures. User-facing changes belong in
|
|
4
|
+
[CHANGELOG.md](CHANGELOG.md) and [GitHub Releases](https://github.com/lambdadb/lambdadb-cli/releases).
|
|
5
|
+
Record execution results in the relevant PR and release, following
|
|
6
|
+
[Recording verification results](CONTRIBUTING.md#recording-verification-results).
|
|
109
7
|
|
|
110
8
|
## Branch and version policy
|
|
111
9
|
|
|
@@ -214,8 +112,8 @@ compatible. Human summaries are not a scripting interface.
|
|
|
214
112
|
|
|
215
113
|
3. Run the live smoke below against an explicitly designated development project.
|
|
216
114
|
Record commit, SDK version, endpoint/project, command, outcome and cleanup
|
|
217
|
-
without keys or signed URLs. Missing credentials are a blocked
|
|
218
|
-
prerequisite, not a skipped success. PR CI deliberately has no service keys.
|
|
115
|
+
in the release PR without keys or signed URLs. Missing credentials are a blocked
|
|
116
|
+
release prerequisite, not a skipped success. PR CI deliberately has no service keys.
|
|
219
117
|
4. For rc/stable releases, promote to main by reviewed PR with required CI
|
|
220
118
|
checks. The first development bootstrap can use reviewed develop directly.
|
|
221
119
|
Re-run checks/live smoke if release contents changed. Record the verified
|
|
@@ -233,7 +131,8 @@ compatible. Human summaries are not a scripting interface.
|
|
|
233
131
|
7. Verify the workflow, npm version/dist-tag, provenance's repository/commit and
|
|
234
132
|
a clean consumer install of the exact version. Test the installed binary's
|
|
235
133
|
version/help and basic JSON behavior. Confirm a dev/rc release did not move
|
|
236
|
-
`latest`.
|
|
134
|
+
`latest`. Add the verification summary and links to the release PR or GitHub
|
|
135
|
+
Release. Synchronize release-only main changes back to develop by PR.
|
|
237
136
|
|
|
238
137
|
Useful read-only checks after publication:
|
|
239
138
|
|
|
@@ -254,8 +153,8 @@ Trusted Publishers are package-specific. The SDK's configuration does not cover
|
|
|
254
153
|
this CLI. npm requires an existing package before configuring its trust policy.
|
|
255
154
|
See the official [npm trust prerequisites](https://docs.npmjs.com/cli/v11/commands/npm-trust/).
|
|
256
155
|
|
|
257
|
-
This
|
|
258
|
-
|
|
156
|
+
This package is already bootstrapped. The steps below document initial setup;
|
|
157
|
+
do not repeat the bootstrap for an existing package or normal dev builds.
|
|
259
158
|
For a new package, first complete the reviewed release preparation and preflight.
|
|
260
159
|
An authorized npm organization owner performs the bootstrap from the verified
|
|
261
160
|
develop commit, using interactive `npm login` and the account's MFA, then
|
|
@@ -297,45 +196,9 @@ this direct-publish workflow. See [npm Trusted Publishing](https://docs.npmjs.co
|
|
|
297
196
|
|
|
298
197
|
## Explicit live smoke
|
|
299
198
|
|
|
300
|
-
|
|
301
|
-
|
|
302
|
-
|
|
303
|
-
|
|
304
|
-
```sh
|
|
305
|
-
export LAMBDADB_ENDPOINT=https://YOUR_DEV_API_ORIGIN
|
|
306
|
-
export LAMBDADB_PROJECT=YOUR_DEV_PROJECT
|
|
307
|
-
# Inject LAMBDADB_API_KEY securely; no .env file is loaded automatically.
|
|
308
|
-
export LAMBDADB_RUN_LIVE_TESTS=1
|
|
309
|
-
export LAMBDADB_LIVE_CONFIRM_PROJECT="$LAMBDADB_PROJECT"
|
|
310
|
-
npm run test:live
|
|
311
|
-
```
|
|
312
|
-
|
|
313
|
-
The test fails before API calls when required settings are absent. It creates a
|
|
314
|
-
random temporary collection, imports two 3 MiB documents plus a bulk document,
|
|
315
|
-
accepts and verifies metadata for all 49 fixed analyzer presets, checks exact
|
|
316
|
-
facet-only and document+facet buckets, and checks committed query/fetch contents
|
|
317
|
-
with a maximum 300-second observation window per stage. It also checks managed reranking with default criteria, null
|
|
318
|
-
and custom criteria, including final/retrieval scores and status metadata.
|
|
319
|
-
The SDK 0.8.0 extension creates two additional temporary Collections for native
|
|
320
|
-
embedding-only and legacy true inputs. It exercises actual CLI create/update,
|
|
321
|
-
normalized metadata, document/query embedding generation, ordinary retrieval,
|
|
322
|
-
Bayesian candidate budgets, default/explicit rerank budgets, null rerank and
|
|
323
|
-
server-side rejection of invalid Bayesian requests. All three Collections are
|
|
324
|
-
cleaned up with subsequent HTTP 404 verification.
|
|
325
|
-
Request timeouts and polling sleeps are capped to the remaining
|
|
326
|
-
budget, and responses at or after the deadline cannot pass. Process termination
|
|
327
|
-
and event-loop scheduling may delay reporting. It reports elapsed time and
|
|
328
|
-
matched-document counts; a timeout is a failed smoke, not proof that accepted
|
|
329
|
-
writes were rejected or lost.
|
|
330
|
-
Collection cleanup uses the SDK because deletion is outside the public CLI MVP.
|
|
331
|
-
The harness verifies deletion with a subsequent HTTP 404.
|
|
332
|
-
Only the test's own collection is eligible for cleanup. An interrupted process
|
|
333
|
-
or failed cleanup may leave that named collection; inspect it manually.
|
|
334
|
-
|
|
335
|
-
This sample does not prove global index readiness, ranking quality, every ref's
|
|
336
|
-
live behavior, or exactly-once writes. A large result alone does not prove the
|
|
337
|
-
service selected docsUrl; mock tests explicitly exercise that transport path.
|
|
338
|
-
Use existing known development tag/alias refs for additional live ref evidence.
|
|
199
|
+
Follow the [live validation procedure](CONTRIBUTING.md#live-validation) and record
|
|
200
|
+
its result in the release PR. It is a release prerequisite; ordinary PR CI has
|
|
201
|
+
no service credentials and cannot establish live readiness.
|
|
339
202
|
|
|
340
203
|
## Homebrew handoff
|
|
341
204
|
|
|
@@ -346,22 +209,7 @@ commit. Development publication does not update Homebrew. After each stable npm
|
|
|
346
209
|
release, follow the [Homebrew maintainer guide](https://github.com/lambdadb/lambdadb-cli/tree/develop/packaging/homebrew#readme) to
|
|
347
210
|
update checksums, pass installation checks and prepare a tap PR. Tap publication
|
|
348
211
|
is separate from this repository's npm workflow; no cross-repository write or
|
|
349
|
-
automatic tap update is configured.
|
|
350
|
-
|
|
351
|
-
(
|
|
352
|
-
|
|
353
|
-
|
|
354
|
-
## Design references
|
|
355
|
-
|
|
356
|
-
- LambdaDB TypeScript client `RELEASING.md` and `publish.yaml`: canonical channels,
|
|
357
|
-
separate live evidence, same tested artifact and OIDC. This CLI intentionally
|
|
358
|
-
publishes dev from develop, while rc/stable releases require main history.
|
|
359
|
-
- LambdaDB migration `.goreleaser.yml`/release workflow and `install.sh`: verify
|
|
360
|
-
installed artifacts and explicit versions. Go binaries, GoReleaser and Docker
|
|
361
|
-
distribution are unnecessary for this npm CLI's first release.
|
|
362
|
-
- [Command Line Interface Guidelines](https://clig.dev/): preserve machine output,
|
|
363
|
-
separate diagnostics, avoid interactive CI requirements, protect secrets and
|
|
364
|
-
test cancellation. The CLI contract suite exercises these boundaries.
|
|
365
|
-
|
|
366
|
-
External documentation checked on 2026-09-18. Consult current npm documentation
|
|
367
|
-
when configuring the package; registry requirements can change independently.
|
|
212
|
+
automatic tap update is configured. Record installation and actual upgrade results
|
|
213
|
+
in the tap PR, including tested platforms and any unperformed checks. Query the
|
|
214
|
+
[public tap](https://github.com/lambdadb/homebrew-tap) for its current formula;
|
|
215
|
+
do not maintain a second release-status history here.
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@functional-systems/lambdadb-cli",
|
|
3
|
-
"version": "0.1.4-dev.
|
|
3
|
+
"version": "0.1.4-dev.18",
|
|
4
4
|
"private": false,
|
|
5
5
|
"description": "Project-scoped LambdaDB command-line tools",
|
|
6
6
|
"type": "module",
|
|
@@ -12,7 +12,6 @@
|
|
|
12
12
|
"src",
|
|
13
13
|
"README.md",
|
|
14
14
|
"DESIGN.md",
|
|
15
|
-
"VALIDATION.md",
|
|
16
15
|
"CONTRIBUTING.md",
|
|
17
16
|
"examples",
|
|
18
17
|
"RELEASING.md",
|
|
@@ -33,7 +32,7 @@
|
|
|
33
32
|
"test:live": "npm run build && node --test test/live/*.test.mjs"
|
|
34
33
|
},
|
|
35
34
|
"dependencies": {
|
|
36
|
-
"@functional-systems/lambdadb": "0.8.
|
|
35
|
+
"@functional-systems/lambdadb": "0.8.1",
|
|
37
36
|
"commander": "14.0.3"
|
|
38
37
|
},
|
|
39
38
|
"devDependencies": {
|
|
@@ -53,5 +52,5 @@
|
|
|
53
52
|
"url": "https://github.com/lambdadb/lambdadb-cli/issues"
|
|
54
53
|
},
|
|
55
54
|
"license": "Apache-2.0",
|
|
56
|
-
"gitHead": "
|
|
55
|
+
"gitHead": "6e4c83681b49eac1c5c5f34f1f87f11aff78d84a"
|
|
57
56
|
}
|