You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The goal is to replicate anafero-cli package functionality in the browser.
Full browser-based wrapper
Currently, the CLI accepts a repository root in the filesystem. It does the following: 1) resolve configuration, 2) fetch, compile & instantiate adapters, 3) invoke anafero’s process() function, providing it requisite utilities to access object data from repository in local filesystem (current dir) and to write resulting artefacts to local filesystem.
We can do the same in the browser (mostly in worker). The difference is that the user would specify a source repository URL & credentials, instead of filesystem we’d use whatever fastest in-browser storage (possibly DuckDB or SQLite, or maybe we can try LightningFS or similar), and adapters would be instantiated using async blob import()s.
Minimal browser-based wrapper
After #37, build system would optionally accept a source and adapter configuration without a Git repository. A simple browser wrapper would not care about working with Git, but instead would accept a dragged file or directory. Configuration can be provided by the user through a simple GUI (since it cannot be resolved). The browser-based wrapper would still need to handle fetching, compiling and instantiating adapters (likely using async blob imports).
After build
After build, we can initiate a download from the browser, or even create an iframe and allow to browse the site.
The text was updated successfully, but these errors were encountered:
The goal is to replicate
anafero-cli
package functionality in the browser.Full browser-based wrapper
Currently, the CLI accepts a repository root in the filesystem. It does the following: 1) resolve configuration, 2) fetch, compile & instantiate adapters, 3) invoke anafero’s process() function, providing it requisite utilities to access object data from repository in local filesystem (current dir) and to write resulting artefacts to local filesystem.
We can do the same in the browser (mostly in worker). The difference is that the user would specify a source repository URL & credentials, instead of filesystem we’d use whatever fastest in-browser storage (possibly DuckDB or SQLite, or maybe we can try LightningFS or similar), and adapters would be instantiated using async blob
import()
s.Minimal browser-based wrapper
After #37, build system would optionally accept a source and adapter configuration without a Git repository. A simple browser wrapper would not care about working with Git, but instead would accept a dragged file or directory. Configuration can be provided by the user through a simple GUI (since it cannot be resolved). The browser-based wrapper would still need to handle fetching, compiling and instantiating adapters (likely using async blob imports).
After build
After build, we can initiate a download from the browser, or even create an iframe and allow to browse the site.
The text was updated successfully, but these errors were encountered: