spin.systems generator and driver
Project description
quill
quill is the "driver" for spin.systems, and packaged on PyPi as ql.
A more detailed description of the package namespace can be found below, but for everyday usage the commands needed are:
import quill as ql
ql.ssm.check_manifest() # Check all component repos' status
ql.ssm.repos_df # print the summary dataframe
ql.fold.wire.standup() # build or rebuild any 'wire' module MMD documents as HTML pages
ql.fold.cut.standup() # build or rebuild any 'cut' module Jinja templates as HTML pages
ql.remote_push_manifest() # Add/commit/push any dirty repos, triggering CI build
The last step will also rebuild any sites which have dirty repos (to avoid content partially falling behind, when for instance the templates for a site are changed).
Structure
ql⠶scan: Read.mmdfilesscan⠶lever: Parse.mmdfile format
ql⠶manifest: Readspin.systemsconfigurationql⠶fold: Manage*.spin.systemssubdomainsfold⠶address: Manage subdomain address shorthandfold⠶wire: Emit HTML websites from.mmdfilesfold⠶cut: Emit HTML websites from.htmlJinja templates and.mdfiles
Helper CLI
Note: for reasons unknown, the
defoptCLI only works on 3.10+
Installation adds a ql command (for deployment) and cyl (for local preview).
These two commands handle repo-internal and -external output of generated sites respectively
(cyl currently only for the fold.cut module, not fold.wire,
so not yet sufficient to replace in the CI pipeline).
usage: ql [-h] [-d [DOMAINS_LIST ...]] [--incremental] [-n] [-r] [-w] [-v]
[-g] [--internal | --no-internal] [--version]
Configure input filtering and output display.
options:
-h, --help show this help message and exit
-d [DOMAINS_LIST ...], --domains-list [DOMAINS_LIST ...]
Names of the subdomains to build.
(default: None)
--incremental compute MD5 checksums for each of the generated files.
(default: False)
-n, --no-render Dry run mode.
(default: False)
-r, --recheck Only regenerate files whose input checksum has changed since the last
incremental run.
(default: False)
-w, --watch Rebuild the domains continuously (incompatible with incremental/recheck).
(default: False)
-v, --verbose Whether to announce what tasks are being carried out.
(default: False)
-g, --gitlab-ci Source the manifest repos (i.e. git pull them) and then stash
the changes after building, switch to the www branch, pop the
stash and push the changes (i.e. publish the website content).
(default: False)
--internal, --no-internal
Whether to build sites internal (ql) or external (cyl,
not including 'fold.wire') to the repo.
(default: True)
--version show program's version number and exit
So for example:
cyl -d pore
Populates ~/spin/cyl/pore with the cut module's portion of pore.spin.systems
Incremental builds, rechecking, and watching
Towards the goal of implementing an incremental build system on CI, quill has 2 extra modes:
- an incremental build mode: where in addition to building the entire site (or domain list),
MD5 checksums are computed for each of the generated files. In this mode, the build auditer is activated,
and stores a log as it generates the templates. Use the
-iflag when you want to recreate the inventory of template MD5 hashes.
To run a build for the same domain as above, in incremental mode, run:
cyl -d pore -i
- a recheck build mode: a subset of incremental builds in which only files whose input checksum
has changed since the last incremental run are regenerated. This makes the build substantially faster
as there is a reduced scope of templates to transform. Use the
-rflag (alongside-i) when you want to double check only the files in the template hash inventory and only regenerate what differs on disk.
To run a recheck afterwards, and only build any files that changed in between builds, run:
cyl -d pore -i -r
- a watch build mode: incompatible with incremental/recheck, this is for continuously rebuilding
(incremental mode is designed for repeated use, like on an incremental build system). Use the
-wflag when you want to have the page rebuild automatically in the background.
Watching is incompatible with incremental mode, but rebuilds following changes to a file.
The corresponding command to the ones above would be:
cyl -d pore -w
Usage memo
- Requires directory of static site repositories at
location specified in
spin.ini(by default this is as a sibling directory../ss/), containing a filemanifest.mmdbeginning with the apex domain and immediately followed by a subdomain list.
Below is a demo of accessing the spin.systems manifest file
(ssm) which implements the MMD class, subclassing Doc,
which parses the file contents in a structured way (specifically,
as a list of colon-separated values).
>>> from pprint import pprint
>>> from quill import ssm
>>> import pandas as pd
>>> ssm
Parsed MMD file (Document of 1 block, containing 1 list)
>>> ssm.list
Headered list with 16 items
>>> pprint(ssm.list.all_nodes)
[-spin.systems:spin-systems:master www:,
-:cal:qu-cal:master www,
-,:log:spin-log:master www,
-,:conf:qu-conf:master www,
-,:pore:qu-pore:master www,
-,:ocu:naiveoculus:master www,
-,:arc:appendens:master www,
-,:qrx:qu-arx:master www,
-,:erg:spin-erg:master www,
-,:opt:spin-opt:master www,
-,:poll:qu-poll:master www,
-,:arb:spin-arb:master www,
-,:reed:qu-reed:master www,
-,:noto:qu-noto:master www,
-,:plot:qu-plot:master www,
-,:doc:spin-doc:master www,
-,:labs:qu-labs:master www]
Similarly ssm.list.nodes gives just the subdomains, and ssm.list.header gives the main site
domain.
>>> ssm.list.header
-spin.systems:spin-systems:
>>> ssm.list.header.parts
['spin.systems', 'spin-systems', 'master www']
>>> ssm.list.nodes[0].parts
['cal', 'qu-cal', 'master www']
>>> ssm.list.nodes[1].parts
['log', 'spin-log', 'master www']
>>> ssm.list.nodes[2].parts
['conf', 'qu-conf', 'master www']
>>> pprint(ssm.all_parts)
[['spin.systems', 'spin-systems', 'master www'],
['cal', 'qu-cal', 'master www'],
['log', 'spin-log', 'master www'],
['conf', 'qu-conf', 'master www'],
['pore', 'qu-pore', 'master www'],
['ocu', 'naiveoculus', 'master www'],
['arc', 'appendens', 'master www'],
['qrx', 'qu-arx', 'master www'],
['erg', 'spin-erg', 'master www'],
['opt', 'spin-opt', 'master www'],
['poll', 'qu-poll', 'master www'],
['arb', 'spin-arb', 'master www'],
['reed', 'qu-reed', 'master www'],
['noto', 'qu-noto', 'master www'],
['plot', 'qu-plot', 'master www'],
['doc', 'spin-doc', 'master www'],
['labs', 'qu-labs', 'master www']]
>>> ssm.as_df()
domain repo_name branches
0 spin.systems spin-systems master www
1 cal qu-cal master www
2 log spin-log master www
3 conf qu-conf master www
4 pore qu-pore master www
5 ocu naiveoculus master www
6 arc appendens master www
7 qrx qu-arx master www
8 erg spin-erg master www
9 opt spin-opt master www
10 poll qu-poll master www
11 arb spin-arb master www
12 reed qu-reed master www
13 noto qu-noto master www
14 plot qu-plot master www
15 doc spin-doc master www
16 labs qu-labs master www
This is just for review purposes currently, and any further info can be added as long as all the lines ("nodes") have the same number of colon-separated values.
The manifest is parsed in manifest⠶parsing by
parse_man_node which is wrapped into ssm.repos,
which will print out the repos' git addresses for each CNAME domain or subdomain:
('spin.systems', 'git@gitlab.com:spin-systems/spin-systems.gitlab.io.git')
('cal', 'git@gitlab.com:qu-cal/qu-cal.gitlab.io.git')
('log', 'git@gitlab.com:spin-log/spin-log.gitlab.io.git')
('conf', 'git@gitlab.com:qu-conf/qu-conf.gitlab.io.git')
('pore', 'git@gitlab.com:qu-pore/qu-pore.gitlab.io.git')
('ocu', 'git@gitlab.com:naiveoculus/naiveoculus.gitlab.io.git')
('arc', 'git@gitlab.com:appendens/appendens.gitlab.io.git')
('qrx', 'git@gitlab.com:qu-arx/qu-arx.gitlab.io.git')
('erg', 'git@gitlab.com:spin-erg/spin-erg.gitlab.io.git')
('opt', 'git@gitlab.com:spin-opt/spin-opt.gitlab.io.git')
('poll', 'git@gitlab.com:qu-poll/qu-poll.gitlab.io.git')
('arb', 'git@gitlab.com:spin-arb/spin-arb.gitlab.io.git')
('reed', 'git@gitlab.com:qu-reed/qu-reed.gitlab.io.git')
('noto', 'git@gitlab.com:qu-noto/qu-noto.gitlab.io.git')
('plot', 'git@gitlab.com:qu-plot/qu-plot.gitlab.io.git')
('doc', 'git@gitlab.com:spin-doc/spin-doc.gitlab.io.git')
('labs', 'git@gitlab.com:qu-labs/qu-labs.gitlab.io.git')
as well as a DataFrame which is modified by the ql.ssm.check_manifest() to include 'live' views on
the repos (note that this method takes a add_before_check=True argument, which controls whether
git add --all is run on each repo to check if it's 'dirty').
>>> ssm.repos_df
domain repo_name branches git_url
0 spin.systems spin-systems master www git@gitlab.com:spin-systems/spin-systems.gitlab.io.git
1 cal qu-cal master www git@gitlab.com:qu-cal/qu-cal.gitlab.io.git
2 log spin-log master www git@gitlab.com:spin-log/spin-log.gitlab.io.git
3 conf qu-conf master www git@gitlab.com:qu-conf/qu-conf.gitlab.io.git
4 pore qu-pore master www git@gitlab.com:qu-pore/qu-pore.gitlab.io.git
5 ocu naiveoculus master www git@gitlab.com:naiveoculus/naiveoculus.gitlab.io.git
6 arc appendens master www git@gitlab.com:appendens/appendens.gitlab.io.git
7 qrx qu-arx master www git@gitlab.com:qu-arx/qu-arx.gitlab.io.git
8 erg spin-erg master www git@gitlab.com:spin-erg/spin-erg.gitlab.io.git
9 opt spin-opt master www git@gitlab.com:spin-opt/spin-opt.gitlab.io.git
10 poll qu-poll master www git@gitlab.com:qu-poll/qu-poll.gitlab.io.git
11 arb spin-arb master www git@gitlab.com:spin-arb/spin-arb.gitlab.io.git
12 reed qu-reed master www git@gitlab.com:qu-reed/qu-reed.gitlab.io.git
13 noto qu-noto master www git@gitlab.com:qu-noto/qu-noto.gitlab.io.git
14 plot qu-plot master www git@gitlab.com:qu-plot/qu-plot.gitlab.io.git
15 doc spin-doc master www git@gitlab.com:spin-doc/spin-doc.gitlab.io.git
16 labs qu-labs master www git@gitlab.com:qu-labs/qu-labs.gitlab.io.git
>>> ssm.check_manifest()
domain repo_name branches git_url branch local clean
0 spin.systems spin-systems master www git@gitlab.com:spin-systems/spin-systems.gitlab.io.git www True True
1 cal qu-cal master www git@gitlab.com:qu-cal/qu-cal.gitlab.io.git www True True
2 log spin-log master www git@gitlab.com:spin-log/spin-log.gitlab.io.git www True True
3 conf qu-conf master www git@gitlab.com:qu-conf/qu-conf.gitlab.io.git www True True
4 pore qu-pore master www git@gitlab.com:qu-pore/qu-pore.gitlab.io.git www True True
5 ocu naiveoculus master www git@gitlab.com:naiveoculus/naiveoculus.gitlab.io.git www True True
6 arc appendens master www git@gitlab.com:appendens/appendens.gitlab.io.git www True True
7 qrx qu-arx master www git@gitlab.com:qu-arx/qu-arx.gitlab.io.git www True True
8 erg spin-erg master www git@gitlab.com:spin-erg/spin-erg.gitlab.io.git www True True
9 opt spin-opt master www git@gitlab.com:spin-opt/spin-opt.gitlab.io.git www True True
10 poll qu-poll master www git@gitlab.com:qu-poll/qu-poll.gitlab.io.git master True True
11 arb spin-arb master www git@gitlab.com:spin-arb/spin-arb.gitlab.io.git www True True
12 reed qu-reed master www git@gitlab.com:qu-reed/qu-reed.gitlab.io.git www True True
13 noto qu-noto master www git@gitlab.com:qu-noto/qu-noto.gitlab.io.git www True True
14 plot qu-plot master www git@gitlab.com:qu-plot/qu-plot.gitlab.io.git www True True
15 doc spin-doc master www git@gitlab.com:spin-doc/spin-doc.gitlab.io.git www True True
16 labs qu-labs master www git@gitlab.com:qu-labs/qu-labs.gitlab.io.git www True True
In this example, the poll repo is on the master branch, and the rest are on the www (web deploy) branch.
Obviously this can then be used to clone the repositories locally
(or address them for any other git-related task)
- One nice feature of GitLab (which at the time of writing GitHub doesn't provide to my knowledge) is that these repos can all be private, and only the static site will be hosted publicly (specified in 'Settings' > 'General')
To clone a given repo (testing has all been with SSH URLs), there is the ql.clone() function,
and subsequently the namespace can be refreshed to reflect the new addition (this is
done automatically within the clone function).
>>> ql.ns
{}
>>> ql.clone(ql.ssm.repos_df.git_url[0], "spin.systems")
Cloning into 'spin.systems'...
remote: Enumerating objects: 6, done.
remote: Counting objects: 100% (6/6), done.
remote: Compressing objects: 100% (6/6), done.
remote: Total 236 (delta 1), reused 0 (delta 0), pack-reused 230
Receiving objects: 100% (236/236), 34.16 KiB | 210.00 KiB/s, done.
Resolving deltas: 100% (123/123), done.
>>> ql.ns
{'spin.systems': 'https://gitlab.com/spin-systems/spin-systems.gitlab.io'}
Lastly, the entire manifest of repos can be sourced from ssm and cloned into the
ns_path directory. This is done on CI to build each site when a change takes place
in one of the source repos or the quill repo (the 'engine').
ql.source_manifest()
For now, if the directory named as the domain entry of the row in the ssm.repos_df
table exists, it will simply not touch it. If it doesn't exist, it will try to clone it.
Et voila the namespace now contains all the repos (stored in the sibling ss directory)
>>> pprint(ql.ns)
{'arb': 'https://gitlab.com/spin-arb/spin-arb.gitlab.io',
'arc': 'https://gitlab.com/appendens/appendens.gitlab.io',
'cal': 'https://gitlab.com/qu-cal/qu-cal.gitlab.io',
'conf': 'https://gitlab.com/qu-conf/qu-conf.gitlab.io',
'doc': 'https://gitlab.com/spin-doc/spin-doc.gitlab.io',
'erg': 'https://gitlab.com/spin-erg/spin-erg.gitlab.io',
'labs': 'https://gitlab.com/qu-labs/qu-labs.gitlab.io',
'log': 'https://gitlab.com/spin-log/spin-log.gitlab.io',
'noto': 'https://gitlab.com/qu-noto/qu-noto.gitlab.io',
'ocu': 'https://gitlab.com/naiveoculus/naiveoculus.gitlab.io',
'opt': 'https://gitlab.com/spin-opt/spin-opt.gitlab.io',
'plot': 'https://gitlab.com/qu-plot/qu-plot.gitlab.io',
'poll': 'https://gitlab.com/qu-poll/qu-poll.gitlab.io',
'pore': 'https://gitlab.com/qu-pore/qu-pore.gitlab.io',
'qrx': 'https://gitlab.com/qu-arx/qu-arx.gitlab.io',
'reed': 'https://gitlab.com/qu-reed/qu-reed.gitlab.io',
'spin.systems': 'https://gitlab.com/spin-systems/spin-systems.gitlab.io'}
At the end of source_manifest, the ssm.repos_df DataFrame is updated with a column local
indicating whether each domain in the manifest is now in the ns namespace (i.e. whether a
local repo has been created), via the check_manifest method which ssm's MMD class inherits
from the Doc class.
- This
update_manifestmethod will be expanded to supplement therepos_dfDataFrame with other information worth knowing to do with thegitstatus of the repo in question, for those which are locally available. This ensures no unnecessary computation is done before the extra information is needed.
The next thing we can do (having established that these are now cloned locally) is to read the CI YAML as the 'layout' for each site, checking they're valid (according to the reference on YAML configs for GitLab Pages)
>>> manifests = ql.yaml_manifests(as_dicts=False)
>>> for k,m in manifests.items(): print(k, end="\t"); pprint(m)
spin.systems SiteCI: PagesJob: {Stage.Deploy, Script <subdirectory='site'>, Artifacts <paths=['public']>, Only <branch=['www']>}
cal SiteCI: PagesJob: {Stage.Deploy, Script <subdirectory='site'>, Artifacts <paths=['public']>, Only <branch=['www']>}
log SiteCI: PagesJob: {Stage.Deploy, Script <subdirectory='site'>, Artifacts <paths=['public']>, Only <branch=['www']>}
conf SiteCI: PagesJob: {Stage.Deploy, Script <subdirectory='site'>, Artifacts <paths=['public']>, Only <branch=['www']>}
pore SiteCI: PagesJob: {Stage.Deploy, Script <subdirectory='site'>, Artifacts <paths=['public']>, Only <branch=['www']>}
ocu SiteCI: PagesJob: {Stage.Deploy, Script <subdirectory='site'>, Artifacts <paths=['public']>, Only <branch=['www']>}
arc SiteCI: PagesJob: {Stage.Deploy, Script <subdirectory='site'>, Artifacts <paths=['public']>, Only <branch=['www']>}
qrx SiteCI: PagesJob: {Stage.Deploy, Script <subdirectory='site'>, Artifacts <paths=['public']>, Only <branch=['www']>}
erg SiteCI: PagesJob: {Stage.Deploy, Script <subdirectory='site'>, Artifacts <paths=['public']>, Only <branch=['www']>}
opt SiteCI: PagesJob: {Stage.Deploy, Script <subdirectory='site'>, Artifacts <paths=['public']>, Only <branch=['www']>}
poll SiteCI: PagesJob: {Stage.Deploy, Script <subdirectory='site'>, Artifacts <paths=['public']>, Only <branch=['www']>}
arb SiteCI: PagesJob: {Stage.Deploy, Script <subdirectory='site'>, Artifacts <paths=['public']>, Only <branch=['www']>}
reed SiteCI: PagesJob: {Stage.Deploy, Script <subdirectory='site'>, Artifacts <paths=['public']>, Only <branch=['www']>}
noto SiteCI: PagesJob: {Stage.Deploy, Script <subdirectory='site'>, Artifacts <paths=['public']>, Only <branch=['www']>}
plot SiteCI: PagesJob: {Stage.Deploy, Script <subdirectory='site'>, Artifacts <paths=['public']>, Only <branch=['www']>}
doc SiteCI: PagesJob: {Stage.Deploy, Script <subdirectory='site'>, Artifacts <paths=['public']>, Only <branch=['www']>}
labs SiteCI: PagesJob: {Stage.Deploy, Script <subdirectory='site'>, Artifacts <paths=['public']>, Only <branch=['www']>}
- Obviously these could all be made more elaborate for less trivial scripts, this is a first draft showing basic functionality and to enforce/check standardisation across all repos
- This setup is deliberately brittle, so any changes will need to be validated in the
fold.yaml_utilmodule, and so the library itself incorporates testing (simpleassertstatements based on a clear expected implementation)
The build directory can be set with change_build_dir, and using this I set all of the repos to
build from "site" (but this can be changed at a later date):
for d in ql.alias_df.domain: ql.change_build_dir(d, "site")
⇣
Moved build path for log from 'site' --> /home/louis/spin/ss/log/site
Created build path for ocu at /home/louis/spin/ss/ocu/site
Moved build path for arc from 'site' --> /home/louis/spin/ss/arc/site
Created build path for erg at /home/louis/spin/ss/erg/site
Created build path for opt at /home/louis/spin/ss/opt/site
Created build path for arb at /home/louis/spin/ss/arb/site
Created build path for doc at /home/louis/spin/ss/doc/site
Created build path for cal at /home/louis/spin/ss/cal/site
Moved build path for conf from 'docs' --> /home/louis/spin/ss/conf/site
Created build path for pore at /home/louis/spin/ss/pore/site
Created build path for qrx at /home/louis/spin/ss/qrx/site
Created build path for poll at /home/louis/spin/ss/poll/site
Created build path for reed at /home/louis/spin/ss/reed/site
Created build path for noto at /home/louis/spin/ss/noto/site
Created build path for plot at /home/louis/spin/ss/plot/site
Created build path for labs at /home/louis/spin/ss/labs/site
To commit these changes, I added some more functions to manage the git repos.
ql.ssm.check_manifest() now has a column referring to whether the working tree
is clean or has changes to tracked files not staged for commit.
The output will be something like this example (where the README in arc
was moved):
ql.remote_push_manifest()
⇣
Skipping 'repo_dir=/home/louis/spin/ss/spin.systems' (working tree clean)
Skipping 'repo_dir=/home/louis/spin/ss/cal' (working tree clean)
Skipping 'repo_dir=/home/louis/spin/ss/log' (working tree clean)
Skipping 'repo_dir=/home/louis/spin/ss/conf' (working tree clean)
Skipping 'repo_dir=/home/louis/spin/ss/pore' (working tree clean)
Skipping 'repo_dir=/home/louis/spin/ss/ocu' (working tree clean)
Commit [repo_dir=/home/louis/spin/ss/arc] ⠶ Renamed site/README.md -> README.md
⇢ Pushing ⠶ origin
Skipping 'repo_dir=/home/louis/spin/ss/qrx' (working tree clean)
Skipping 'repo_dir=/home/louis/spin/ss/erg' (working tree clean)
Skipping 'repo_dir=/home/louis/spin/ss/opt' (working tree clean)
Skipping 'repo_dir=/home/louis/spin/ss/poll' (working tree clean)
Skipping 'repo_dir=/home/louis/spin/ss/arb' (working tree clean)
Skipping 'repo_dir=/home/louis/spin/ss/reed' (working tree clean)
Skipping 'repo_dir=/home/louis/spin/ss/noto' (working tree clean)
Skipping 'repo_dir=/home/louis/spin/ss/plot' (working tree clean)
Skipping 'repo_dir=/home/louis/spin/ss/doc' (working tree clean)
Skipping 'repo_dir=/home/louis/spin/ss/labs' (working tree clean)
This function takes a refspec argument which indicates a particular path to add and push in each repo,
a specific_domains argument which can be a string of a single domain, or list of multiple,
and will be evaluated as the list of all domains if left as None.
Running ssm.check_manifest() again is required to update ssm.repos_df.
This repos_df dataframe is also useful for comparing whatever
other properties you might want to check, e.g. which have a README
>>> df = ql.ssm.repos_df
>>> df["has_README"] = ["README.md" in [x.name for x in (ql.ns_path / d).iterdir()] for d in df.domain]
>>> df
domain repo_name branches ... local clean has_README
0 spin.systems spin-systems master www ... True True True
1 cal qu-cal master www ... True True False
2 log spin-log master www ... True True True
3 conf qu-conf master www ... True True True
4 pore qu-pore master www ... True True False
5 ocu naiveoculus master www ... True True True
6 arc appendens master www ... True True True
7 qrx qu-arx master www ... True True False
8 erg spin-erg master www ... True True False
9 opt spin-opt master www ... True True False
10 poll qu-poll master www ... True True False
11 arb spin-arb master www ... True True False
12 reed qu-reed master www ... True True False
13 noto qu-noto master www ... True True False
14 plot qu-plot master www ... True True True
15 doc spin-doc master www ... True True False
16 labs qu-labs master www ... True True False
Obviously this is the kind of thing to then follow up manually, but it helps to have programmatic ways to view the set of directories.
If after reviewing ssm.repos_df you want to push, you can do so
either with an automated commit message or by passing it as the commit_msg
argument to remote_push_manifest (which will reuse the same commit message
if multiple repos are not clean).
- Note you can always manually go in and check the
git diffbeforehand
static, src, and wire
To build all sites with a src and/or static directory, run ql.fold.cut.standup().
This builds templates from the src folder with staticjinja, and the static files
are just copied over (directly under site/).
To build all sites with a wire config, run ql.fold.wire.standup().
More details
In fact, standup returns a dictionary of the domains, though for now
only one domain is in use with wires.
emitters = ql.fold.wire.standup(verbose=False)
emitters
⇣
{'poll': <quill.fold.wire.emitters.WireEmitter object at 0x7f304980eb20>}
Then to push the sites 'live', run ql.remote_push_manifest("Commit message goes here").
Aliases
The "canonical names" displayed in the README are the aliases, which by and large are the same as the domain names.
These provide the titles of the index pages of each site.
I might make more use of these in future, I originally only needed these
as it turns out the description on GitHub/GitLab isn't stored in the git repo
itself (silly as there's a description built in to git?)
Using this let me generate a nice README for the spin.systems superrepo:
spin.systems
spin.systems:
ss
Addresses
"Spin addresses" follow the above "{namespace}⠶{domain|alias}" format, and additionally:
- Some (initially only
∫⠶log) subdomain repos are 'deploy' stage counterparts to local 'dev' stage directories. - Some (initially only
∫⠶log) subdomain repos will be 'addressed' with a date [and a zero-based counter for same-day entries]
>>> example = "∫⠶log⠶20⠶oct⠶25⠶0"
>>> eg_addr = ql.AddressPath(example)
>>> eg_addr
['∫', 'log', '20', 'oct', '25', '0']
The path has been parsed ("strictly" by default) into parts which are 'typed' strings.
pprint(list(map(type, eg_addr)))
⇣
[<class 'ql.src.quill.scan.address.paths.NameSpaceString'>,
<class 'ql.src.quill.scan.address.paths.DomainString'>,
<class 'ql.src.quill.scan.address.paths.YyDigitString'>,
<class 'ql.src.quill.scan.address.paths.mmmString'>,
<class 'ql.src.quill.scan.address.paths.DdDigitString'>,
<class 'ql.src.quill.scan.address.paths.FileIntString'>]
A file path can be obtained from this using interpret_filepath,
which is bound to the class as the filepath property:
>>> eg_addr.filepath
PosixPath('/home/louis/spin/l/20/10oct/25/0_digitalising_spin_addresses.mmd')
>>> eg_addr.filepath.exists()
True
>>> ql.mmd(eg_addr.filepath)
Parsed MMD file (Document of 4 blocks)
This comes in handy when building components of the spin.systems site such as
tap which can then build parts for a particular domain
with
>>> eg_addr = ql.AddressPath.from_parts(domain="poll", ymd=(2021, 2, 17))
>>> eg_addr.filepath
PosixPath('/home/louis/spin/ss/poll/transmission/21/02feb/17')
This gives a simple date-based interface obeying the storage structure of quill, though unlike files, paths to directories in this way may not exist (instead they can be created as needed).
TODO
- A next step could be a class representing the state of the websites [beyond CI], which can
then be cross-referenced against the
repos_df(but the goal is not to entirely Python-ise the site development, just the management of key aspects to do with the version control on disk) - Make a pip installable binary wheel (bdist not currently working with SCM, just sdist)
- Make package capable of downloading missing data files in the event it is being distributed
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
