# Disable automatic selection of full word when using ner.correct recipe

**URL:** <https://support.prodi.gy/t/disable-automatic-selection-of-full-word-when-using-ner-correct-recipe/7106>\
**Category:** Uncategorized\
**Created:** [February 17, 2024, 4:48pm UTC](https://support.prodi.gy/t/disable-automatic-selection-of-full-word-when-using-ner-correct-recipe/7106 "2024-02-17T16:48:30Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![fernandorodriguespro](https://sea2.discourse-cdn.com/flex020/user_avatar/support.prodi.gy/fernandorodriguespro/32/4291_2.png) [@fernandorodriguespro](https://support.prodi.gy/u/fernandorodriguespro)\
**Post date:** [February 17, 2024, 4:48pm UTC](https://support.prodi.gy/t/disable-automatic-selection-of-full-word-when-using-ner-correct-recipe/7106/1 "2024-02-17T16:48:30Z")

</div>

When using prodigy interface to select words, if there is no space the entire string is selected. Somehow I am not able to select characters. How can it be solved?

In the following image, I just want VITAMINA C, but all previous characters are selected automatically, because there is no space between them.

 ![WhatsApp Image 2024-02-15 at 18.05.23](https://us1.discourse-cdn.com/flex020/uploads/prodigy/original/2X/8/86efcd1cdd3b5aec9b854ae84fa449e64d1c13c0.jpeg)

This will affect model prediction performance.

---

<div class="post-metadata">

**Author:** ![magdaaniol](https://sea2.discourse-cdn.com/flex020/user_avatar/support.prodi.gy/magdaaniol/32/2787_2.png) [@magdaaniol](https://support.prodi.gy/u/magdaaniol)\
**Post date:** [February 19, 2024, 3:39pm UTC](https://support.prodi.gy/t/disable-automatic-selection-of-full-word-when-using-ner-correct-recipe/7106/2 "2024-02-19T15:39:27Z")

</div>

Welcome to the forum @fernandorodriguespro ,

In `ner.manual` and `spans.manual` recipes you can add a toggle for switching between character and token-based annotation by calling them with the [`--highlight-chars`](https://prodi.gy/docs/named-entity-recognition#highlight-chars) flag. Be mindful, though, that this won't affect the tokenization so you'll end up with misaligned tokens and spans. In order to train a model afterwards, you'll need a tokenizer that will be able to recreate this same tokenization (See the important note in the docs referred above)

The typical workflow in these cases is to use this `highlight-chars` feature to "record" the issues in the tokenization of your dataset and use them to create [custom tokenization rules](https://spacy.io/usage/linguistic-features#native-tokenizers) for the tokenizer that will be used in the training.  
You can also use the manually retokenized phrases to test the custom tokenizer that you will create.  
From there you might integrate the custom tokenizer in your training pipeline or you might also choose to solve these issues in the preprocessing step (as this is where this problem essentially belongs) - it shouldn't really matter.

---

<div class="post-metadata">

**Author:** ![fernandorodriguespro](https://sea2.discourse-cdn.com/flex020/user_avatar/support.prodi.gy/fernandorodriguespro/32/4291_2.png) [@fernandorodriguespro](https://support.prodi.gy/u/fernandorodriguespro)\
**Post date:** [February 20, 2024, 1:09am UTC](https://support.prodi.gy/t/disable-automatic-selection-of-full-word-when-using-ner-correct-recipe/7106/3 "2024-02-20T01:09:20Z")

</div>

What if I just skip these words?

---

<div class="post-metadata">

**Author:** ![magdaaniol](https://sea2.discourse-cdn.com/flex020/user_avatar/support.prodi.gy/magdaaniol/32/2787_2.png) [@magdaaniol](https://support.prodi.gy/u/magdaaniol)\
**Post date:** [February 23, 2024, 9:32am UTC](https://support.prodi.gy/t/disable-automatic-selection-of-full-word-when-using-ner-correct-recipe/7106/4 "2024-02-23T09:32:27Z")

</div>

If you skip these words from processing, you won't be able to extract any information from it. If you believe that the information you're after appears correctly tokenized in other parts of text and you would be getting enough training examples despite ignoring mistokenized words then it should OK to ignore them.

This [thread](https://support.prodi.gy/t/nlp-span-categorization-compound-words-or-wordsstucktogetherlikethis/7080) is in-depth discussion of such "agglutinations" - it might be of interest to you as well.