Tag Settings
Taglinker scan all the tags in your documents and links them, it will guess any incomplete or broken tags based on your tag name, separators in the name and the minimal length of the tag name.
Therefore, you need to provide your standard separators based on your project tag naming convention. You are allow to specify 2 separators and the minimal lenght of your tag name.
TagLinker supports most of the tag name separators used in the industry. Below, you have a list of the tag settings required by TagLinker.
- Separator 1
- Separator 2 (if any)
- Minimum Length
- Index
- Status
The interface looks like this:
Separators types 1,2
TagLinker let you set 2 separators which is common practice in the industry.The interface looks like this:
You can choose among 5 types of separators in the list box, see below:
In the tag below, you have only one type of separator "-" in blue below.
In the tag below, you have two types of separator "-" in blue and "/" in green below.
Minimum Length
The OCR module of TagLinker will scan the document and ignore any text below the minimum length, this length must be aligned with your tag naming convention. You can choose between 4 to 9 characters.
See the video guide below for tag seetings and lenght:
Language
By default the OCR will try to guess your language however, it is better performance wise to select the language yourself. The OCR supports more than 50 languages wordwide. So far, the OCR was tested with the english language only but feel free to try your own country language and let us know if you encounters any problems.
The "Default" Behavior (No Hint) Automatic Detection: The API will analyze the image and, for each word or block of text, try to determine the most likely language.
Best for Mixed Languages: This is the best approach for a document like your P&ID, which contains both English and Chinese. It allows the API to correctly identify and label each language without you having to separate them or make multiple API calls.
General Case: For most high-quality, clear images with a mix of languages, leaving the language hint empty will give you the best overall results. It prevents a situation where you "force" the API to look for the wrong language.
When a Language Hint Is Useful Low-Quality Images: If the image is blurry, has poor lighting, or the text is small and difficult to read, providing a hint can guide the model to the correct character set. This can significantly reduce "hallucinations" (where the model guesses incorrectly) and improve the accuracy of the recognized characters.
Single-Language Documents: If you know for a fact that the entire document is in a single language, a hint can streamline the process and potentially increase accuracy. For example, if you are processing a batch of scanned documents that are all in German, specifying de can improve the results.
Similar Languages: In cases where two languages share many characters or have similar scripts (e.g., Cyrillic-based languages like Russian and Ukrainian, or Latin-based languages like Spanish and Italian), a hint can help the API differentiate and choose the correct one.
Status
Status parameter is a optional parameter, but can be very useful during the commissioning of your project. You can create a list of status for your tags such as "not installed", "connected", "tested", etc.. and affect a color for each status, any tags on the document having this status will appear with their specific color right on your document giving you the status of all the tags at a glance.
See the video guide about how to set a tag status: