Skip to main content

Google Sheets

Google Sheets for fancy-flow — as four imported, versioned packages, one per runtime. Not vendored source: a copy cannot be upgraded, and third-party APIs change.

Runtime Package Install
Authoring surface (every host) @particle-academy/google-sheets-ui npm install @particle-academy/google-sheets-ui
Node @particle-academy/google-sheets-js npm install @particle-academy/google-sheets-js
PHP 8.4+ particle-academy/google-sheets-php composer require particle-academy/google-sheets-php
Python 3.11+ fancy-google-sheets pip install fancy-google-sheets

The ui package is the editor surface and is React on every host — a PHP or Python project installs it and its own runtime package, and never the js one.

What it costs you

One dependency: @particle-academy/fancy-connector-core (or particle-academy/fancy-connector-core on Composer), which the js and php packages pull in themselves. The Python package has zero runtime dependencies.

No Google Sheets SDK. Plain HTTP, deliberately: a vendor SDK is third-party code subject to the kit's full approval bar, and one per provider is hundreds of dependencies nobody is tracking.

Setting it up

Everything below is generated from provider/manifest.json, so it cannot disagree with what the packages do.

Credentials

A Google Sheets connection holds 4 values.

Two kinds of value, and mixing them up matters. A provider credential is ONE value for the whole installation — an OAuth app's client secret serves every connected account. An account credential is one per connected account. A host that stores the second where it stores the first lets one account's credentials reach another's.

Field Scope Secret Where it comes from
OAuth client ID per installation not secret From Google Cloud Console -> APIs & Services -> Credentials. ONE value for the whole installation, not per connected account.
OAuth client secret per installation secret The client secret for the same OAuth app. One value for the whole installation.
Access token per connected account secret Per connected Google account, and it expires after ONE HOUR. The host refreshes it with the refresh token.
Refresh token per connected account secret Per connected Google account. Google issues one only when the consent request asks for offline access; without it the connection dies within the hour.

Authorising

Google Sheets uses OAuth2 (authorization_code). The package DECLARES the exchange; the HOST performs it — a consent screen needs a browser, a redirect URI and somewhere to persist the result, and all three belong to the host.

The refresh tokens do not rotate: the same one is reusable, so a refresh may safely be retried and may run concurrently. Stated rather than assumed, because the opposite — a provider that spends the token and revokes the grant on a replay — looks identical until it happens.

The estate

Google Sheets has no test estate, and somebody checked. Everything this connector does is real. Use the faker to build against it.

Google has no sandbox for Sheets. A test spreadsheet is a real one in a real Drive, so every append is real -- point this at a scratch sheet, not a production one. The faker is the only way to develop against it without touching a document.

What it can do

Actions

row_append — Google Sheets row

Append a row to a Google Sheet.

POST /v4/spreadsheets/{spreadsheetId}/values/{range}:append · unsafe to replay — a retried durable run does it TWICE

Input Required What it is
spreadsheetId yes The long id from the sheet's URL: docs.google.com/spreadsheets/d/THIS_PART/edit.
range yes Where to append, in A1 notation. Google finds the last row of this range and writes below it.
values yes The cells of the row, comma separated. Left to right from the start of the range.
valueInputOption no USER_ENTERED parses the cells the way typing them would -- so =SUM(A1:A2) becomes a formula and 1/2/26 becomes a date. RAW stores exactly the characters given.
insertDataOption no INSERT_ROWS pushes existing rows down. OVERWRITE writes into them.

Run it before you have credentials

Every operation ships a faker, whether or not Google Sheets has a sandbox. Set a node's mode to fake and it returns the shape Google Sheets actually publishes — the same field names, deterministically — so you can wire the downstream nodes before touching an account, a key, or a network.

This repository is generated

provider/ is the source. Everything under packages/ is emitted from it and must not be hand-edited — CI regenerates and diffs on every push, and the next protocol sync destroys anything it finds. See AGENTS.md.

Two namespaces, which do not match on purpose

The repo is github.com/Fancy-Friends/google-sheets; the packages publish under particle-academy. Nothing derives one from the other — the names come from weaver's friends.json and nowhere else.

Licence

MIT.

Metadata

Release files for fancy-google-sheets 0.3.3

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for fancy-google-sheets 0.3.3
File Size Uploaded
fancy_google_sheets-0.3.3.tar.gz 19.0 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for fancy-google-sheets 0.3.3
File Interpreter ABI Platform
fancy_google_sheets-0.3.3-py3-none-any.whl Python 3 none any Details

Total release size: 37.5 kB

Release files / fancy_google_sheets-0.3.3.tar.gz

Download URL fancy_google_sheets-0.3.3.tar.gz
Size 19.0 kB
Tags Source
SHA-256 checksum
How to use checksums
bf493df1766453c2abdf52c1c48b33894cea31b2f813a6c156d87b191d528e40
BLAKE2b-256 checksum
How to use checksums
0a5c6a01a1209e81db045559737d8add4b906139ea526732dfea115f7c576fe3
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.13.15

Release files / fancy_google_sheets-0.3.3-py3-none-any.whl

Download URL fancy_google_sheets-0.3.3-py3-none-any.whl
Size 18.4 kB
Tags Python 3
SHA-256 checksum
How to use checksums
6fc915a22dd505fe964f4c4fb0314b4a7729cdec68c7550583a63b3850c564dd
BLAKE2b-256 checksum
How to use checksums
24bb51164dca4e228db0a034e817fec75cfc9a1de1184c2ac94a01517a3de494
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.13.15

Release history Release notifications | RSS feed

0.4.0

2 release files

0.3.4

2 release files

This release

0.3.3 This release

2 release files

0.3.2

2 release files

0.3.1

2 release files

0.3.0

2 release files

0.2.0

2 release files

0.1.0

2 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page