1datalad rerun(1) General Commands Manual datalad rerun(1)
2
6 datalad rerun - re-execute previous `datalad run` commands.
7
9 datalad rerun [-h] [--since SINCE] [-d DATASET] [-b NAME] [-m MESSAGE]
10 [--onto base] [--script FILE] [--report] [--assume-
11 ready {inputs|outputs|both}] [--explicit] [-J NJOBS] [--version]
12 [REVISION]
13
17 This will unlock any dataset content that is on record to have been
18 modified by the command in the specified revision. It will then re-ex‐
19 ecute the command in the recorded path (if it was inside the dataset).
20 Afterwards, all modifications will be saved.
21 Report mode
23 When called with --report, this command reports information about what
24 would be re-executed as a series of records. There will be a record for
25 each revision in the specified revision range. Each of these will have
26 one of the following "rerun_action" values:
27 - run: the revision has a recorded command that would be re-executed -
29 skip-or-pick: the revision does not have a recorded command and would
30 be either skipped or cherry picked - merge: the revision is a merge
31 commit and a corresponding merge would
32 be made
33 The decision to skip rather than cherry pick a revision is based on
35 whether the revision would be reachable from HEAD at the time of execu‐
36 tion.
37 In addition, when a starting point other than HEAD is specified, there
39 is a rerun_action value "checkout", in which case the record includes
40 information about the revision the would be checked out before rerun‐
41 ning any commands.
42 NOTE Currently the "onto" feature only sets the working tree of the
44 current dataset to a previous state. The working trees of any
45 subdatasets remain unchanged.
46 Examples
48 Re-execute the command from the previous commit::
49 % datalad rerun
51 Re-execute any commands in the last five commits::
53 % datalad rerun --since=HEAD~5
55 Do the same as above, but re-execute the commands on top of HEAD~5 in a
57 detached state::
58 % datalad rerun --onto= --since=HEAD~5
60 Re-execute all previous commands and compare the old and new results::
62 % # on master branch
64 % datalad rerun --branch=verify --since=
65 % # now on verify branch
66 % datalad diff --revision=master..
67 % git log --oneline --left-right --cherry-pick master...
68
71 REVISION
72 rerun command(s) in REVISION. By default, the command from this
73 commit will be executed, but --since can be used to construct a
74 revision range. The default value is like "HEAD" but resolves to
75 the main branch when on an adjusted branch. Constraints: value
76 must be a string or value must be NONE
77 -h, --help, --help-np
80 show this help message. --help-np forcefully disables the use of
81 a pager for displaying the help message
82 --since SINCE
84 If SINCE is a commit-ish, the commands from all commits that are
85 reachable from `revision` but not SINCE will be re-executed (in
86 other words, the commands in git log SINCE..REVISION). If SINCE
87 is an empty string, it is set to the parent of the first commit
88 that contains a recorded command (i.e., all commands in git log
89 REVISION will be re-executed). Constraints: value must be a
90 string or value must be NONE
91 -d DATASET, --dataset DATASET
93 specify the dataset from which to rerun a recorded command. If
94 no dataset is given, an attempt is made to identify the dataset
95 based on the current working directory. If a dataset is given,
96 the command will be executed in the root directory of this
97 dataset. Constraints: Value must be a Dataset or a valid identi‐
98 fier of a Dataset (e.g. a path) or value must be NONE
99 -b NAME, --branch NAME
101 create and checkout this branch before rerunning the commands.
102 Constraints: value must be a string or value must be NONE
103 -m MESSAGE, --message MESSAGE
105 use MESSAGE for the reran commit rather than the recorded commit
106 message. In the case of a multi-commit rerun, all the reran com‐
107 mits will have this message. Constraints: value must be a string
108 or value must be NONE
109 --onto base
111 start point for rerunning the commands. If not specified, com‐
112 mands are executed at HEAD. This option can be used to specify
113 an alternative start point, which will be checked out with the
114 branch name specified by --branch or in a detached state other‐
115 wise. As a special case, an empty value for this option means
116 the parent of the first run commit in the specified revision
117 list. Constraints: value must be a string or value must be NONE
118 --script FILE
120 extract the commands into FILE rather than rerunning. Use - to
121 write to stdout instead. This option implies --report. Con‐
122 straints: value must be a string or value must be NONE
123 --report
125 Don't actually re-execute anything, just display what would be
126 done. Note: If you give this option, you most likely want to set
127 --output-format to 'json' or 'json_pp'.
128 --assume-ready {inputs|outputs|both}
130 Assume that inputs do not need to be retrieved and/or outputs do
131 not need to unlocked or removed before running the command. This
132 option allows you to avoid the expense of these preparation
133 steps if you know that they are unnecessary. Note that this op‐
134 tion also affects any additional outputs that are automatically
135 inferred based on inspecting changed files in the run commit.
136 Constraints: value must be one of ('inputs', 'outputs', 'both')
137 --explicit
139 Consider the specification of inputs and outputs in the run
140 record to be explicit. Don't warn if the repository is dirty,
141 and only save modifications to the outputs from the original
142 record. Note that when several run commits are specified, this
143 applies to every one. Care should also be taken when using --on‐
144 to because checking out a new HEAD can easily fail when the
145 working tree has modifications.
146 -J NJOBS, --jobs NJOBS
148 how many parallel jobs (where possible) to use. "auto" corre‐
149 sponds to the number defined by 'datalad.runtime.max-annex-jobs'
150 configuration item NOTE: This option can only parallelize input
151 retrieval (get) and output recording (save). DataLad does NOT
152 parallelize your scripts for you. Constraints: value must be
153 convertible to type 'int' or value must be NONE or value must be
154 one of ('auto',)
155 --version
157 show the module and its version which provides the command
158
160 datalad is developed by The DataLad Team and Contributors <team@datal‐
161 ad.org>.
162datalad rerun 0.19.3 2023-08-11 datalad rerun(1)