diff --git a/converters/Makefile b/converters/Makefile index d963b78583d0..645b3b83065f 100644 --- a/converters/Makefile +++ b/converters/Makefile @@ -1,189 +1,190 @@ COMMENT = Character code converters SUBDIR += R-cran-RJSONIO SUBDIR += R-cran-arabic2kansuji SUBDIR += R-cran-base64enc SUBDIR += R-cran-jsonlite SUBDIR += R-cran-rjson SUBDIR += aish SUBDIR += ascii2binary SUBDIR += asr10 SUBDIR += base64 SUBDIR += base91 SUBDIR += bibtexconv SUBDIR += bmfdec SUBDIR += bsdconv SUBDIR += btoa SUBDIR += chmview SUBDIR += cl-babel SUBDIR += cl-babel-sbcl SUBDIR += cmios9 SUBDIR += cocot SUBDIR += convmv SUBDIR += dos2ansi SUBDIR += dos2unix SUBDIR += dosunix SUBDIR += drg2sbg SUBDIR += dumpasn1 SUBDIR += enca SUBDIR += fconv SUBDIR += fix-mime-charset SUBDIR += fondu SUBDIR += fribidi SUBDIR += gbase SUBDIR += gbsdconv SUBDIR += hs-aeson-pretty SUBDIR += hs-dhall-bash SUBDIR += hs-dhall-json SUBDIR += hs-dhall-yaml SUBDIR += ical2html SUBDIR += iconv SUBDIR += iconv-extra SUBDIR += iconv-rfc1345 SUBDIR += ish SUBDIR += libb64 SUBDIR += libiconv SUBDIR += libticonv SUBDIR += libutf-8 SUBDIR += linux-c7-fribidi SUBDIR += linux-rl9-fribidi SUBDIR += lua-iconv SUBDIR += lua-json SUBDIR += mimelib SUBDIR += mmencode SUBDIR += mpack SUBDIR += nomyso SUBDIR += o3read SUBDIR += ocaml-base64 SUBDIR += ocaml-jsonm SUBDIR += osm2mp SUBDIR += osm2pgrouting SUBDIR += osm2pgsql SUBDIR += p5-Bencode SUBDIR += p5-Boulder SUBDIR += p5-Convert-ASCII-Armour SUBDIR += p5-Convert-ASN1 SUBDIR += p5-Convert-BER SUBDIR += p5-Convert-Base32 SUBDIR += p5-Convert-Bencode SUBDIR += p5-Convert-Bencode_XS SUBDIR += p5-Convert-BinHex SUBDIR += p5-Convert-Binary-C SUBDIR += p5-Convert-IBM390 SUBDIR += p5-Convert-Moji SUBDIR += p5-Convert-Morse SUBDIR += p5-Convert-NLS_DATE_FORMAT SUBDIR += p5-Convert-PEM SUBDIR += p5-Convert-Recode SUBDIR += p5-Convert-TNEF SUBDIR += p5-Convert-UU SUBDIR += p5-Convert-UUlib SUBDIR += p5-Cpanel-JSON-XS SUBDIR += p5-Cstools SUBDIR += p5-Data-AMF SUBDIR += p5-Encode SUBDIR += p5-Encode-Base32-Crockford SUBDIR += p5-Encode-Base32-GMP SUBDIR += p5-Encode-Base58-GMP SUBDIR += p5-Encode-Detect SUBDIR += p5-Encode-Detect-CJK SUBDIR += p5-Encode-DoubleEncodedUTF8 SUBDIR += p5-Encode-IMAPUTF7 SUBDIR += p5-Encode-JIS2K SUBDIR += p5-Encode-Locale SUBDIR += p5-Encode-Punycode SUBDIR += p5-Encode-compat SUBDIR += p5-Encoding-BER SUBDIR += p5-Encoding-FixLatin SUBDIR += p5-Format-Human-Bytes SUBDIR += p5-JSON SUBDIR += p5-JSON-Any SUBDIR += p5-JSON-DWIW SUBDIR += p5-JSON-MaybeXS SUBDIR += p5-JSON-PP SUBDIR += p5-JSON-Tiny SUBDIR += p5-JSON-XS SUBDIR += p5-JSON-XS-VersionOneAndTwo SUBDIR += p5-LaTeXML SUBDIR += p5-MIME-Base32 SUBDIR += p5-MIME-Base64 SUBDIR += p5-MIME-Base64-URLSafe SUBDIR += p5-Net-IDN-Nameprep SUBDIR += p5-Number-Nary SUBDIR += p5-Number-RecordLocator SUBDIR += p5-PerlIO-via-Unidecode SUBDIR += p5-Sereal SUBDIR += p5-Sereal-Decoder SUBDIR += p5-Sereal-Encoder SUBDIR += p5-Storable-AMF SUBDIR += p5-String-Multibyte SUBDIR += p5-Text-Bidi SUBDIR += p5-Text-Iconv SUBDIR += p5-Text-Unidecode SUBDIR += p5-URI-UTF8-Punycode SUBDIR += p5-Unicode-IMAPUtf7 SUBDIR += p5-Unicode-Map SUBDIR += p5-Unicode-Map8 SUBDIR += p5-Unicode-MapUTF8 SUBDIR += p5-Unicode-RecursiveDowngrade SUBDIR += p5-Unicode-String SUBDIR += p5-Unicode-Stringprep SUBDIR += p5-Unicode-UTF8simple SUBDIR += p5-WAP-wbxml SUBDIR += p5-XML-WBXML SUBDIR += p5-bsdconv SUBDIR += p5-plutil SUBDIR += pdf2djvu SUBDIR += pear-Services_JSON SUBDIR += pecl-igbinary SUBDIR += php-bsdconv SUBDIR += php81-iconv SUBDIR += php81-mbstring SUBDIR += php82-iconv SUBDIR += php82-mbstring SUBDIR += php83-iconv SUBDIR += php83-mbstring SUBDIR += php84-iconv SUBDIR += php84-mbstring SUBDIR += psiconv SUBDIR += py-bencode.py SUBDIR += py-bencoder SUBDIR += py-better-bencode SUBDIR += py-bincopy SUBDIR += py-bsdconv SUBDIR += py-gotenberg-client SUBDIR += py-mammoth + SUBDIR += py-markitdown SUBDIR += py-rencode SUBDIR += py-svglib SUBDIR += py-text-unidecode SUBDIR += py-tlv8 SUBDIR += py-unidecode SUBDIR += py-webencodings SUBDIR += py-zfec SUBDIR += rcctools SUBDIR += recode SUBDIR += rubygem-base32 SUBDIR += rubygem-base64 SUBDIR += rubygem-iconv SUBDIR += rubygem-json-ld SUBDIR += rubygem-json-ld-preloaded SUBDIR += rubygem-po_to_json SUBDIR += rubygem-ruby-bsdconv SUBDIR += rubygem-url_safe_base64 SUBDIR += rubygem-xdr SUBDIR += showkey SUBDIR += simdutf SUBDIR += ta2as SUBDIR += tnef SUBDIR += trans SUBDIR += unix2dos SUBDIR += uudeview SUBDIR += uulib SUBDIR += wkhtmltopdf SUBDIR += xdeview SUBDIR += xml2c SUBDIR += yj SUBDIR += yj-bruceadams SUBDIR += ytnef .include diff --git a/converters/py-markitdown/Makefile b/converters/py-markitdown/Makefile new file mode 100644 index 000000000000..a9ce7a689d57 --- /dev/null +++ b/converters/py-markitdown/Makefile @@ -0,0 +1,27 @@ +PORTNAME= markitdown +DISTVERSION= 0.0.1a3 +CATEGORIES= converters python +MASTER_SITES= PYPI +PKGNAMEPREFIX= ${PYTHON_PKGNAMEPREFIX} + +MAINTAINER= wen@FreeBSD.org +COMMENT= Utility tool for converting various files to Markdown +WWW= https://pypi.org/project/tlv8/ + +LICENSE= APACHE20 + +BUILD_DEPENDS= ${PYTHON_PKGNAMEPREFIX}hatchling>=0:devel/py-hatchling@${PY_FLAVOR} +RUN_DEPENDS= ${PYTHON_PKGNAMEPREFIX}mammoth>=0:converters/py-mammoth@${PY_FLAVOR} \ + ${PYTHON_PKGNAMEPREFIX}markdownify>=0:textproc/py-markdownify@${PY_FLAVOR} \ + ${PYTHON_PKGNAMEPREFIX}pandas>=0:math/py-pandas@${PY_FLAVOR} \ + ${PYTHON_PKGNAMEPREFIX}pdfminer.six>=0:textproc/py-pdfminer.six@${PY_FLAVOR} \ + ${PYTHON_PKGNAMEPREFIX}python-pptx>=0:textproc/py-python-pptx@${PY_FLAVOR} \ + ${PYTHON_PKGNAMEPREFIX}puremagic>=0:sysutils/py-puremagic@${PY_FLAVOR} \ + ${PYTHON_PKGNAMEPREFIX}requests>=0:www/py-requests@${PY_FLAVOR} + +USES= python +USE_PYTHON= autoplist pep517 + +NO_ARCH= yes + +.include diff --git a/converters/py-markitdown/distinfo b/converters/py-markitdown/distinfo new file mode 100644 index 000000000000..a69065a058ef --- /dev/null +++ b/converters/py-markitdown/distinfo @@ -0,0 +1,3 @@ +TIMESTAMP = 1734654122 +SHA256 (markitdown-0.0.1a3.tar.gz) = f6c8f5f7f5541e91c6c535218318968fefd71e2a6faa0eb782b3492e04cd023d +SIZE (markitdown-0.0.1a3.tar.gz) = 16073 diff --git a/converters/py-markitdown/pkg-descr b/converters/py-markitdown/pkg-descr new file mode 100644 index 000000000000..8871cf0e5603 --- /dev/null +++ b/converters/py-markitdown/pkg-descr @@ -0,0 +1,13 @@ +MarkItDown library is a utility tool for converting various files to Markdown +(e.g., for indexing, text analysis, etc.) + +It presently supports: + *PDF (.pdf) + *PowerPoint (.pptx) + *Word (.docx) + *Excel (.xlsx) + *Images (EXIF metadata, and OCR) + *Audio (EXIF metadata, and speech transcription) + *HTML (special handling of Wikipedia, etc.) + *Various other text-based formats (csv, json, xml, etc.) + *ZIP (Iterates over contents and converts each file)