|
| 1 | +#!/bin/bash -eu |
| 2 | +# Copyright 2025 Google LLC |
| 3 | +# |
| 4 | +# Licensed under the Apache License, Version 2.0 (the "License"); |
| 5 | +# you may not use this file except in compliance with the License. |
| 6 | +# You may obtain a copy of the License at |
| 7 | +# |
| 8 | +# http://www.apache.org/licenses/LICENSE-2.0 |
| 9 | +# |
| 10 | +# Unless required by applicable law or agreed to in writing, software |
| 11 | +# distributed under the License is distributed on an "AS IS" BASIS, |
| 12 | +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| 13 | +# See the License for the specific language governing permissions and |
| 14 | +# limitations under the License. |
| 15 | +# |
| 16 | +################################################################################ |
| 17 | + |
| 18 | +mkdir ${SRC}/seeds |
| 19 | +#This packages the unit test files based on file extension from within the Tika project |
| 20 | +#we could also pull in other seeds from other parser projects. |
| 21 | + |
| 22 | +find ${SRC}/project-parent/tika -name "*-webm.noext" -print0 | xargs -0 zip -u ${SRC}/seeds/AudioVideoParsersFuzzer_seed_corpus.zip |
| 23 | +find ${SRC}/project-parent/tika -name "*-mkv.noext" -print0 | xargs -0 zip -u ${SRC}/seeds/AudioVideoParsersFuzzer_seed_corpus.zip |
| 24 | +find ${SRC}/project-parent/tika -name "*.aif" -print0 | xargs -0 zip -u ${SRC}/seeds/AudioVideoParsersFuzzer_seed_corpus.zip |
| 25 | +find ${SRC}/project-parent/tika -name "*.au" -print0 | xargs -0 zip -u ${SRC}/seeds/AudioVideoParsersFuzzer_seed_corpus.zip |
| 26 | +find ${SRC}/project-parent/tika -name "*.flv" -print0 | xargs -0 zip -u ${SRC}/seeds/AudioVideoParsersFuzzer_seed_corpus.zip |
| 27 | +find ${SRC}/project-parent/tika -name "*.m4a" -print0 | xargs -0 zip -u ${SRC}/seeds/AudioVideoParsersFuzzer_seed_corpus.zip |
| 28 | +find ${SRC}/project-parent/tika -name "*.mkv" -print0 | xargs -0 zip -u ${SRC}/seeds/AudioVideoParsersFuzzer_seed_corpus.zip |
| 29 | +find ${SRC}/project-parent/tika -name "*.mp3" -print0 | xargs -0 zip -u ${SRC}/seeds/AudioVideoParsersFuzzer_seed_corpus.zip |
| 30 | +find ${SRC}/project-parent/tika -name "*.wav" -print0 | xargs -0 zip -u ${SRC}/seeds/AudioVideoParsersFuzzer_seed_corpus.zip |
| 31 | + |
| 32 | + |
| 33 | +find ${SRC}/project-parent/tika -name "*.Z" -print0 | xargs -0 zip -u ${SRC}/seeds/CompressorParserFuzzer_seed_corpus.zip |
| 34 | +find ${SRC}/project-parent/tika -name "*.bz2" -print0 | xargs -0 zip -u ${SRC}/seeds/CompressorParserFuzzer_seed_corpus.zip |
| 35 | +find ${SRC}/project-parent/tika -name "*.gz" -print0 | xargs -0 zip -u ${SRC}/seeds/CompressorParserFuzzer_seed_corpus.zip |
| 36 | +find ${SRC}/project-parent/tika -name "*.tbz2" -print0 | xargs -0 zip -u ${SRC}/seeds/CompressorParserFuzzer_seed_corpus.zip |
| 37 | +find ${SRC}/project-parent/tika -name "*.tgz" -print0 | xargs -0 zip -u ${SRC}/seeds/CompressorParserFuzzer_seed_corpus.zip |
| 38 | +find ${SRC}/project-parent/tika -name "*.zst" -print0 | xargs -0 zip -u ${SRC}/seeds/CompressorParserFuzzer_seed_corpus.zip |
| 39 | + |
| 40 | +find ${SRC}/project-parent/tika -name "*.html" -print0 | xargs -0 zip ${SRC}/seeds/HtmlParserFuzzer_seed_corpus.zip |
| 41 | + |
| 42 | +find ${SRC}/project-parent/tika -name "*.avif" -print0 | xargs -0 zip -u ${SRC}/seeds/ImageParsersFuzzer_seed_corpus.zip |
| 43 | +find ${SRC}/project-parent/tika -name "*.bmp" -print0 | xargs -0 zip -u ${SRC}/seeds/ImageParsersFuzzer_seed_corpus.zip |
| 44 | +find ${SRC}/project-parent/tika -name "*.bpg" -print0 | xargs -0 zip -u ${SRC}/seeds/ImageParsersFuzzer_seed_corpus.zip |
| 45 | +find ${SRC}/project-parent/tika -name "*.gif" -print0 | xargs -0 zip -u ${SRC}/seeds/ImageParsersFuzzer_seed_corpus.zip |
| 46 | +find ${SRC}/project-parent/tika -name "*.heic" -print0 | xargs -0 zip -u ${SRC}/seeds/ImageParsersFuzzer_seed_corpus.zip |
| 47 | +find ${SRC}/project-parent/tika -name "*.icns" -print0 | xargs -0 zip -u ${SRC}/seeds/ImageParsersFuzzer_seed_corpus.zip |
| 48 | +find ${SRC}/project-parent/tika -name "*.jp2" -print0 | xargs -0 zip -u ${SRC}/seeds/ImageParsersFuzzer_seed_corpus.zip |
| 49 | +find ${SRC}/project-parent/tika -name "*.jb2" -print0 | xargs -0 zip -u ${SRC}/seeds/ImageParsersFuzzer_seed_corpus.zip |
| 50 | +find ${SRC}/project-parent/tika -name "*.jpg" -print0 | xargs -0 zip -u ${SRC}/seeds/ImageParsersFuzzer_seed_corpus.zip |
| 51 | +find ${SRC}/project-parent/tika -name "*.jxl" -print0 | xargs -0 zip -u ${SRC}/seeds/ImageParsersFuzzer_seed_corpus.zip |
| 52 | +find ${SRC}/project-parent/tika -name "*.png" -print0 | xargs -0 zip -u ${SRC}/seeds/ImageParsersFuzzer_seed_corpus.zip |
| 53 | +find ${SRC}/project-parent/tika -name "*.psd" -print0 | xargs -0 zip -u ${SRC}/seeds/ImageParsersFuzzer_seed_corpus.zip |
| 54 | +find ${SRC}/project-parent/tika -name "*.tif" -print0 | xargs -0 zip -u ${SRC}/seeds/ImageParsersFuzzer_seed_corpus.zip |
| 55 | +find ${SRC}/project-parent/tika -name "*.webp" -print0 | xargs -0 zip -u ${SRC}/seeds/ImageParsersFuzzer_seed_corpus.zip |
| 56 | + |
| 57 | + |
| 58 | +find ${SRC}/project-parent/tika -name "*.mdb" -print0 | xargs -0 zip ${SRC}/seeds/JackcessParserFuzzer_seed_corpus.zip |
| 59 | +find ${SRC}/project-parent/tika -name "*.accdb" -print0 | xargs -0 zip ${SRC}/seeds/JackcessParserFuzzer_seed_corpus.zip |
| 60 | + |
| 61 | +find ${SRC}/project-parent/tika -name "*.one" -print0 | xargs -0 zip ${SRC}/seeds/OneNoteParserFuzzer_seed_corpus.zip |
| 62 | + |
| 63 | +#we could get more seeds by cloning POI |
| 64 | +find ${SRC}/project-parent/tika -name "*.msg" -print0 | xargs -0 zip -u ${SRC}/seeds/OfficeParserFuzzer_seed_corpus.zip |
| 65 | +find ${SRC}/project-parent/tika -name "*.doc" -print0 | xargs -0 zip -u ${SRC}/seeds/OfficeParserFuzzer_seed_corpus.zip |
| 66 | +find ${SRC}/project-parent/tika -name "*.ppt" -print0 | xargs -0 zip -u ${SRC}/seeds/OfficeParserFuzzer_seed_corpus.zip |
| 67 | +find ${SRC}/project-parent/tika -name "*.xls" -print0 | xargs -0 zip -u ${SRC}/seeds/OfficeParserFuzzer_seed_corpus.zip |
| 68 | + |
| 69 | +find ${SRC}/project-parent/tika -name "*.docm" -print0 | xargs -0 zip -u ${SRC}/seeds/OOXMLParserFuzzer_seed_corpus.zip |
| 70 | +find ${SRC}/project-parent/tika -name "*.docx" -print0 | xargs -0 zip -u ${SRC}/seeds/OOXMLParserFuzzer_seed_corpus.zip |
| 71 | +find ${SRC}/project-parent/tika -name "*.pptm" -print0 | xargs -0 zip -u ${SRC}/seeds/OOXMLParserFuzzer_seed_corpus.zip |
| 72 | +find ${SRC}/project-parent/tika -name "*.pptx" -print0 | xargs -0 zip -u ${SRC}/seeds/OOXMLParserFuzzer_seed_corpus.zip |
| 73 | +find ${SRC}/project-parent/tika -name "*.xlsm" -print0 | xargs -0 zip -u ${SRC}/seeds/OOXMLParserFuzzer_seed_corpus.zip |
| 74 | +find ${SRC}/project-parent/tika -name "*.xlsx" -print0 | xargs -0 zip -u ${SRC}/seeds/OOXMLParserFuzzer_seed_corpus.zip |
| 75 | + |
| 76 | +find ${SRC}/project-parent/tika -name "*.7z" -print0 | xargs -0 zip -u ${SRC}/seeds/PackageParserFuzzer_seed_corpus.zip |
| 77 | +find ${SRC}/project-parent/tika -name "*.ar" -print0 | xargs -0 zip -u ${SRC}/seeds/PackageParserFuzzer_seed_corpus.zip |
| 78 | +find ${SRC}/project-parent/tika -name "*.jar" -print0 | xargs -0 zip -u ${SRC}/seeds/PackageParserFuzzer_seed_corpus.zip |
| 79 | +find ${SRC}/project-parent/tika -name "*.rar" -print0 | xargs -0 zip -u ${SRC}/seeds/PackageParserFuzzer_seed_corpus.zip |
| 80 | +find ${SRC}/project-parent/tika -name "*.tar" -print0 | xargs -0 zip -u ${SRC}/seeds/PackageParserFuzzer_seed_corpus.zip |
| 81 | +find ${SRC}/project-parent/tika -name "*.zip" -print0 | xargs -0 zip -u ${SRC}/seeds/PackageParserFuzzer_seed_corpus.zip |
| 82 | +find ${SRC}/project-parent/tika -name "*.zlib" -print0 | xargs -0 zip -u ${SRC}/seeds/PackageParserFuzzer_seed_corpus.zip |
| 83 | + |
| 84 | + |
| 85 | +#we could get more seeds by cloning PDFBox or...? |
| 86 | +find ${SRC}/project-parent/tika -name "*.pdf" -print0 | xargs -0 zip ${SRC}/seeds/PDFParserFuzzer_seed_corpus.zip |
| 87 | + |
| 88 | +find ${SRC}/project-parent/tika -name "*.eml" -print0 | xargs -0 zip ${SRC}/seeds/RFC822ParserFuzzer_seed_corpus.zip |
| 89 | + |
| 90 | +find ${SRC}/project-parent/tika -name "*.rtf" -print0 | xargs -0 zip ${SRC}/seeds/RTFParserFuzzer_seed_corpus.zip |
| 91 | + |
| 92 | +find ${SRC}/project-parent/tika -name "*.txt" -print0 | xargs -0 zip ${SRC}/seeds/TextAndCSVParserFuzzer_seed_corpus.zip |
| 93 | +find ${SRC}/project-parent/tika -name "*.tsv" -print0 | xargs -0 zip -u ${SRC}/seeds/TextAndCSVParserFuzzer_seed_corpus.zip |
| 94 | +find ${SRC}/project-parent/tika -name "*.csv" -print0 | xargs -0 zip -u ${SRC}/seeds/TextAndCSVParserFuzzer_seed_corpus.zip |
| 95 | + |
| 96 | +find ${SRC}/project-parent/tika -name "*.xml" -print0 | xargs -0 zip ${SRC}/seeds/XMLReaderUtilsFuzzer_seed_corpus.zip |
| 97 | + |
| 98 | +find ${SRC}/project-parent/tika -path '*/test-documents/*' -type f | xargs -n1 -d '\n' zip ${SRC}/seeds/AutoDetectParserFuzzer_seed_corpus.zip |
| 99 | + |
| 100 | +cp ${SRC}/seeds/*_seed_corpus.zip ${OUT}/ |
| 101 | + |
0 commit comments