📢March 2023 release
This release introduces enhanced flexibility and accuracy, featuring individual feature enabling/disabling, manual page management control, improved enrichments with callbacks, more language options
Last updated
This release introduces enhanced flexibility and accuracy, featuring individual feature enabling/disabling, manual page management control, improved enrichments with callbacks, more language options
Last updated
We have improved the flexibility of our release channels based on your feedback. Previously, the setup did not allow for enough customization. Now, we have made changes to the way feature channels and flags work. Features will still go through different channels based on their stability and maturity, but you can now enable or disable features individually for each project.
For instance, if you require a specific feature that is only available in the experimental channel, you can now enable that feature without getting all the other additional features. This new feature customization allows for greater flexibility and tailored experiences for each project.
We value your opinion and feedback on this new feature, so let us know what you think!
We've made some significant improvements to the user interface for setting up enrichments, including better grouping of settings to make it easier to use. We've also made changes to the Value Types section, which lets you choose how users interact with enrichments, by allowing you to define tables yourself. Another new feature is the ability to use user input to update your own database through callbacks. With these powerful new tools, you can get the most out of manual interventions from your users!
In cases where Metamaze encounters unknown or ambiguous languages, there are a few options available. When dealing with unknown languages, you can either add the language to the list of active project languages, force the system to choose from the active project languages, proceed to validation, or treat the language as "Other". For ambiguous languages, you can choose the language of the first page of your upload, or proceed to validation.
The latest update includes the OCR score in the confidence score. This helps to measure the accuracy of the extracted text, allowing you to consider the quality of the text extraction process when determining the confidence level of an entity's prediction. However, it's important to note that adding the OCR score might lead to a lower overall confidence score.
We are excited to offer you the flexibility to control manual page management. You can easily enable or disable this feature according to your needs. If you choose to disable manual page management, you won't be able to make manual adjustments during an upload if the pages were not properly split (if page management was enabled). However, if you prefer to have more control over page management, you can enable this feature and manually intervene as needed.
With this update, any duplicate emails that have already been processed will be ignored. This is achieved by analyzing several key pieces of information from the email, including the subject line, sender's address, recipients' addresses, body content, and timestamp of when it was sent. By considering these factors, we can ensure that no time is wasted on processing redundant emails. This feature will help to streamline your project's email management and save valuable time and resources.
