docs: explain custom analyzer docker builds

CHANGELOG.md

@@ -4,6 +4,10 @@ All notable changes to this project will be documented in this file.
 
 ## [unreleased]
 
+### General
+#### Changed
+- Documented how to build custom `presidio-analyzer` Docker images with alternative analyzer, NLP, and recognizer-registry YAML files, including multilingual configuration checks and common startup warnings. Fixes [#1663](https://github.com/microsoft/presidio/issues/1663).
+
 ### Anonymizer
 #### Fixed
 - Custom operator `validate()` no longer calls the user-supplied lambda with a dummy `"PII"` value. Previously, stateful lambdas (e.g. those accumulating a token-to-original-value map for de-anonymization) would receive a spurious invocation during validation, inserting a junk entry (`{"TOKEN_1": "PII"}`) into the map and skewing all subsequent token counters. The return-type contract is now enforced in `operate()` when the lambda runs on real data. Fixes [#2024](https://github.com/microsoft/presidio/issues/2024).
@@ -809,4 +813,3 @@ New endpoint for deanonymizing encrypted entities by the anonymizer.
 [2.2.23]: https://github.com/microsoft/presidio/compare/2.2.2...2.2.23
 [2.2.2]: https://github.com/microsoft/presidio/compare/2.2.1...2.2.2
 [2.2.1]: https://github.com/microsoft/presidio/compare/2.2.0...2.2.1
-

docs/installation.md

@@ -96,10 +96,10 @@ To download the Presidio Docker containers, run the following command:
 
     This requires Docker to be installed. [Download Docker](https://docs.docker.com/get-docker/).
 
-### For PII anonymization in text
-
-For PII detection and anonymization in text, the `presidio-analyzer`
-and `presidio-anonymizer` modules are required.
+### For PII anonymization in text
+
+For PII detection and anonymization in text, the `presidio-analyzer`
+and `presidio-anonymizer` modules are required.
 
 ```sh
 # Download Docker images
@@ -109,12 +109,72 @@ docker pull mcr.microsoft.com/presidio-anonymizer
 # Run containers with default ports
 docker run -d -p 5002:3000 mcr.microsoft.com/presidio-analyzer:latest
 
-docker run -d -p 5001:3000 mcr.microsoft.com/presidio-anonymizer:latest
-```
-
-### For PII redaction in images
-
-For PII detection in images, the `presidio-image-redactor` is required.
+docker run -d -p 5001:3000 mcr.microsoft.com/presidio-anonymizer:latest
+```
+
+#### Building a custom `presidio-analyzer` image
+
+The published analyzer image ships with the default English-only YAML files:
+
+- `presidio_analyzer/conf/default_analyzer.yaml`
+- `presidio_analyzer/conf/default.yaml`
+- `presidio_analyzer/conf/default_recognizers.yaml`
+
+To add more languages or enable a different recognizer mix, create custom copies
+of those files inside the `presidio-analyzer/` build context and point the Docker
+build to them with build arguments.
+
+For example:
+
+```sh
+cp presidio-analyzer/presidio_analyzer/conf/default_analyzer.yaml \
+  presidio-analyzer/presidio_analyzer/conf/custom_analyzer.yaml
+cp presidio-analyzer/presidio_analyzer/conf/default.yaml \
+  presidio-analyzer/presidio_analyzer/conf/custom_nlp.yaml
+cp presidio-analyzer/presidio_analyzer/conf/default_recognizers.yaml \
+  presidio-analyzer/presidio_analyzer/conf/custom_recognizers.yaml
+```
+
+Then update the copies as needed:
+
+- Add every runtime language to `supported_languages` in both
+  `custom_analyzer.yaml` and `custom_recognizers.yaml`.
+- Configure the NLP models for those languages in `custom_nlp.yaml`.
+- Enable or add recognizers that should run for the new languages.
+
+Build the image with the custom file paths:
+
+```sh
+docker build ./presidio-analyzer \
+  -t presidio/presidio-analyzer-custom \
+  --build-arg ANALYZER_CONF_FILE=presidio_analyzer/conf/custom_analyzer.yaml \
+  --build-arg NLP_CONF_FILE=presidio_analyzer/conf/custom_nlp.yaml \
+  --build-arg RECOGNIZER_REGISTRY_CONF_FILE=presidio_analyzer/conf/custom_recognizers.yaml
+```
+
+Run it the same way as the default image:
+
+```sh
+docker run -d -p 5002:3000 presidio/presidio-analyzer-custom
+```
+
+!!! note "Important configuration checks"
+
+    - `supported_languages` must match between the analyzer and recognizer-registry YAML files.
+    - The Docker build installs the NLP models declared in `NLP_CONF_FILE` (or in the analyzer file if you use a single combined config), so adding more or larger models increases build time and image size.
+    - If the container logs warnings such as `NLP recognizer ... is not in the list of recognizers for language ...`, check that the recognizer registry still includes the NLP recognizer for that language (for example `SpacyRecognizer`) and that the same language code appears in both YAML files.
+    - If you are experimenting with many languages at once, start with a smaller subset first. Downloading and loading several large NLP models in one image can significantly increase memory usage during build and startup.
+
+For more background on the YAML structure, see:
+
+- [Analyzer Engine Provider](analyzer/analyzer_engine_provider.md)
+- [PII detection in different languages](analyzer/languages.md)
+- [Customizing NLP models](analyzer/customizing_nlp_models.md)
+- [Recognizer registry configuration](analyzer/recognizer_registry_provider.md)
+
+### For PII redaction in images
+
+For PII detection in images, the `presidio-image-redactor` is required.
 
 ```sh
 # Download Docker image