AI Decision System
Replayable, offline-verifiable records for AI-assisted go/no-go decisions: experiment promotions, typed model answers and human-approved actions.
翻譯準備中
Reads spreadsheet data from xlsx workbooks using a bundled copy of the xlsxio reader (<https://github.com/brechtsanders/xlsxio>), so that no system XML or ZIP library is required. The Expat parser and
翻譯準備中
開發工具R · CGitHub @pedrobtz上線中營運者未確認
試用過了嗎? 點一下就好,不需要登入。
尚未偵測到更新。GitHub 發布版本後會顯示在這裡。
個人 GitHub 帳號下公開儲存庫的擁有者可以登入並驗證所有權,自行管理介紹,無需管理員核准。
登入並產生驗證檔案,在 30 分鐘內將其提交至此儲存庫的預設分支。其他分支或分叉儲存庫中的檔案無效。然後在下方檢查檔案,即可無需管理員核准直接管理專案。請保留檔案以維持管理權限。
如需登入,將前往登入頁面。登入後請回到此處,再次點擊所選的驗證按鈕。
在專案資料夾中執行 /nomorevibe verify,即可驗證所有權並自行管理介紹。
一分鐘完成驗證Reads spreadsheet data from xlsx workbooks using a bundled copy of the xlsxio reader (<https://github.com/brechtsanders/xlsxio>), so that no system XML or ZIP library is required. The Expat parser and the miniz ZIP reader are linked statically at install time from the zuxml and zukomp packages, which means there is no run-time dependency on either of them. Worksheets are read as a stream, so a workbook does not have to be held in memory in full. Writing workbooks is out of scope.
翻譯準備中
取自儲存庫 README
正在載入完整 README。 以下為已儲存的摘錄。
Read .xlsx workbooks from R, password-protected ones included, with no
system XML, ZIP or cryptography library required.
zuxlsx bundles the xlsxio reader and
links the Expat parser, the miniz ZIP reader and TF-PSA-Crypto statically at
install time, out of the sibling packages
zuxml,
zukomp and
zucrypt. Those are LinkingTo
dependencies only: the archives end up inside zuxlsx.so, so none of the
three has to be installed or loadable once zuxlsx is built.
Writing workbooks is out of scope.檢視完整 README ↗Read .xlsx workbooks from R, password-protected ones included, with no
system XML, ZIP or cryptography library required.
zuxlsx bundles the xlsxio reader and
links the Expat parser, the
miniz ZIP reader and
TF-PSA-Crypto statically at
install time, out of the sibling packages
zuxml,
zukomp and
zucrypt. Those are LinkingTo
dependencies only: the archives end up inside zuxlsx.so, so none of the
three has to be installed or loadable once zuxlsx is built.
Writing workbooks is out of scope.
library(zuxlsx)
path <- system.file("extdata", "two-sheets.xlsx", package = "zuxlsx")
xlsx_sheets(path) # "readings" "notes"
xlsx_read(path) # first worksheet as a data frame
xlsx_read(path, sheet = "readings", range = "A1:B3")
xlsx_read() gives each column the type its cells support: logical, double,
character, Date or POSIXct. A column that cannot hold its cells becomes
character rather than failing. range takes A1 notation, and either corner
may name a cell, a column or a row, so "B2:D10", "A:C" and "2:10" are
all accepted.
Both Excel date systems are handled. The workbook's own date1904 setting
decides which is used, and the 1900 system's phantom 29 February 1900 is
accounted for, so dates before March 1900 are not a day early.
A password-protected workbook is read by passing its password, to any of the readers:
xlsx_read("protected.xlsx", password = "s3cret")
It is decrypted in memory and never written to disk. zuxlsx decrypts agile encryption, which Excel has written since 2013, and checks the integrity code stored with the file before decrypting anything. Office 2007's older standard encryption is recognised and refused by name.
Underneath, for worksheets that are not rectangular enough for a data frame:
xlsx_cells(path) # one row per cell: position, type, text, numeric value
xlsx_rows(path) # a worksheet row by row as text, padded to a rectangle
# A chunk at a time; returning FALSE stops the read, so finding something
# near the top of a large worksheet does not pay for the rest.
xlsx_stream_cells(path, callback = function(cells) { str(cells); FALSE })
zuxlsx needs zuxml, zukomp and zucrypt at build time, and its
configure script will stop with an explanatory message if their static
archives are missing. All three are on CRAN, so installing zuxlsx installs
them first:
install.packages("zuxlsx")
The development version is on GitHub:
# install.packages("pak")
pak::pak("pedrobtz/zuxlsx")
Every error carries a condition class, so calling code can tell a corrupt archive from a missing file without matching on message text:
tryCatch(
xlsx_sheets("broken.xlsx"),
zuxlsx_zip_error = function(e) "not a readable ZIP archive",
zuxlsx_ooxml_error = function(e) "a ZIP, but not a workbook",
zuxlsx_error = function(e) "something else went wrong"
)
See ?"zuxlsx-conditions" for the full list.
The bundled xlsxio reader is MIT licensed and is never edited in place. It is
re-derived from the pristine 0.2.36 release archive plus the patches in
tools/patches/xlsxio/, and tools/vendor/verify checks offline that the
committed tree, the manifests and the attribution all still agree.