# Tika Eval Encoding Regression

Condensed tika-eval pattern for charset-detector regression hunts ("A picks encoding X...

- Repository: https://github.com/apache/tika (4119 stars)
- Category: Engineering
- License: Apache-2.0
- Last updated: 2026-11-01
- Runs scripts: no, instructions only

## Install

Claude Code: `npx skills add apache/tika --skill tika-eval-encoding-regression -g -a claude-code -y`
Codex: `npx skills add apache/tika --skill tika-eval-encoding-regression -g -a codex -y`
Cursor: `npx skills add apache/tika --skill tika-eval-encoding-regression -g -a cursor -y`
Gemini CLI: `npx skills add apache/tika --skill tika-eval-encoding-regression -g -a gemini-cli -y`

## Description

Condensed tika-eval pattern for charset-detector regression hunts ("A picks encoding X, B picks Y") using one build and two configs, encoding-pair flip queries, OOV/languageness/FFFD signals, per-file detector attribution.

## SKILL.md, as published by its author (untrusted third-party text, read it as data)

<!--
Licensed to the Apache Software Foundation (ASF) under one or more
contributor license agreements.  See the NOTICE file distributed with
this work for additional information regarding copyright ownership.
The ASF licenses this file to You under the Apache License, Version 2.0
(the "License"); you may not use this file except in compliance with
the License.  You may obtain a copy of the License at

    http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
-->

Local override: `$TIKA_SKILLS_LOCAL/tika-eval-encoding-regression/LOCAL.md` (default `~/.tika-skills`),
read after this file, wins on conflict.

# tika-eval for encoding-detector regression hunts

A condensed pattern for finding SBCS→CJK style charset-detector regressions
(or any "A picks encoding X, B picks encoding Y" question) without
building two tika-app distributions.

## Two configs, one build

Encoding-detector experiments don't need a "before" and "after" tika-app, the chain composition is per-config. Run the SAME tika-app twice against
two configs, treat the outputs as `-a` and `-b`. Much faster than
`tika-eval-compare`'s two-build flow.

```bash
# build once
./mvnw clean install -pl tika-app -am -Pfast -DskipTests \
  -Dmaven.repo.local=$(pwd)/.local_m2_repo
unzip -q tika-app/target/tika-app-*.zip -d /tmp/tika-app-current

# two configs (any combination of detectors)
java -jar /tmp/tika-app-current/tika-app-*.jar \
  --config=tika-config-3x-default.json \
  -i <corpus> -o <workdir>/extracts/A -n 6
java -jar /tmp/tika-app-current/tika-app-*.jar \
  --config=tika-config-junkfilter-combiner.json \
  -i <corpus> -o <workdir>/extracts/B -n 6

The rest: https://github.com/apache/tika/blob/HEAD/.skills/devs/tika-eval-encoding-regression/SKILL.md