> ## Documentation Index
> Fetch the complete documentation index at: https://docs.mirobody.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Standardization

> 印刷された測定値を、オフラインで比較できる値にする仕組み。コードを付けるか、理由を添えてコードなしで残し、推測はしません。

対象 mirobody
`1.5.10`
· 元ファイル
[`docs/site/standardization.md`](https://github.com/thetahealth/mirobody/blob/b37cd7b24af5251a4364aab30905a7497fc69dcb/docs/site/standardization.md)

> [!NOTE]
>
> このページの日本語版はまだありません。以下は英語の原文です。

Standardization makes readings from different sources comparable when their
meaning and units match. It runs offline and needs no model API key.

## Names, codes and units

* Measurements use the bundled LOINC vocabulary.
* Reported symptoms and diagnoses use separate ICPC-3 lookups.
* Units are normalized to UCUM; conversions require compatible dimensions.
* Multilingual aliases can resolve to the same concept. An unknown or
  ambiguous name can remain unresolved.

Passing the original unit is important: it can affect which code represents
a reading. Similar labels do not guarantee the same clinical meaning.

## Python example

```python theme={null}
from mirobody import resolve

result = resolve("hemoglobin")
print(result.loinc)  # 718-7
```

Use the original value and unit when standardizing a complete reading.
For document extraction and library examples, see the
[quickstart](https://docs.mirobody.ai/ja/quickstart.md).

## Scope

The bundled vocabulary covers selected laboratory and clinical observations,
not every LOINC concept or vendor score. Incompatible units are not silently
converted, and an unresolved name does not receive a guessed nearest code.
You can inspect the corpus identifier through `mirobody.BUNDLE_VERSION`.

Vocabulary coverage, licensing and mapping rules are documented in the
[standardization reference](https://github.com/thetahealth/mirobody/blob/b37cd7b24af5251a4364aab30905a7497fc69dcb/docs/standardization.md).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.