summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorBryan Newbold <bnewbold@robocracy.org>2018-04-15 00:08:17 -0700
committerBryan Newbold <bnewbold@robocracy.org>2018-04-15 00:08:17 -0700
commit60fba1b8651544e0fb7399a82d1aaff949a896c4 (patch)
tree539a547b3f7b0a2e852e4c91fedda902f589a340
parent097479244bcfcd9e016b0f5a842672018c0a49b5 (diff)
downloadfatcat-60fba1b8651544e0fb7399a82d1aaff949a896c4.tar.gz
fatcat-60fba1b8651544e0fb7399a82d1aaff949a896c4.zip
notes/TODO
-rw-r--r--TODO52
-rw-r--r--notes/data_model.md76
-rw-r--r--notes/url_structure.txt1
3 files changed, 110 insertions, 19 deletions
diff --git a/TODO b/TODO
index db059f15..e30c7028 100644
--- a/TODO
+++ b/TODO
@@ -1,20 +1,50 @@
-- tests are picking up config.py instead of using setUp()
+high-level this weekend:
+- test: create "rich" work (all entities; group approve), all via API
+- test: merge two rich works, all via API
+- naive API-based import scripts for: journals (norwegian), orcid, crossref
-helpers/ORM and test
-x files
-x containers
-- citations
-x create from crossref dict (naive)
-- "hydrated" get: release, creator, container
-- populate_random_edit; push edit to live
-- helpers with change logging (hook?)
+helpers/ORM
+- method to "approve" edit_group
+- helpers to deal with edits and edit groups
+
+api
+- expose all models (like work)
+- expose edit_group and editor
+
+tests
+- create "rich" work (all entities) via edit mechanism (incl. approve)
+- modify existing release via edit mechanism (and commit)
+- merge two releases
+- update (via edit) a redirect release
+- merge two works (combining releases)
+- api: try to reuse an accepted edit group
+- api: try to modify an accepted edit
+- api: multiple edits, same entity
review
-- should release point to work? I think not
-- remove 'state' and 'redirect_id' from all revision tables
+- hydrate in files for releases... nested good enough?
+- add a 'live' (or 'immutable') flag to revision tables
+- how to encode proposed redirects? history goes in changelog
+ => proposed_ident_action table, which points to edits
+ => ident in edit as a partial solution (not redirects)
+ => extend edit object to have "to/from" info, and be per-entity
+
+views
+- my edits/groups
+- oldest edits/edit-groups
later:
+- "hydrate" files (and maybe container/authors/refs) in release
+- transclude primary_release in work
- crossref json import script/benchmark
=> maybe both "raw" and string-dedupe?
- public IDs are UUID (sqlite hack?)
+
+## High-Level Priorities
+
+- manual editing of containers and works/releases
+- bulk loading of releases, files, containers
+- accurate auto-matching matching of containers (eg, via ISSN)
+- full database dump and reload
+
diff --git a/notes/data_model.md b/notes/data_model.md
index fe21b533..94716b00 100644
--- a/notes/data_model.md
+++ b/notes/data_model.md
@@ -1,12 +1,20 @@
+Entity list:
+
+ work
+ release
+ creator
+ container
+ file
+
## Cookbook
-To create a new works:
+To create a new work:
login/create account
- match/lookup on first work; if found, modify not create
- TODO: match/lookup on files?
+ match/lookup on first work; check if update actually needed
+ ? match/lookup on files
new edit group (under account; wip)
new edit (under group)
@@ -47,12 +55,30 @@ To edit, eg, a contributor:
Accept edit group:
- for each edit:
- update entity ident state (activate, redirect, delete)
- ? mark revision/edit as immutable
- ? append log/changelog row
+ for each entity type:
+ for each edit:
+ update entity ident state (activate, redirect, delete)
+ append log/changelog row
update edit group state
+
+Bulk/Fase Import Crossref:
+
+ lookup work by identifier; if exists done
+ lookup journals by ISSN
+ lookup authors by ORCID
+ create new work
+ => stub container and authors if we can't find them
+ create new release
+ submit edit group
+
+
+Import Journals (same for authors):
+
+ lookup journal by ISSN
+ create new container
+ submit edit group
+
## Entity Schema
Each entity type has tables:
@@ -68,7 +94,7 @@ Each entity type has tables:
represents change metadata for a single change to one ident
needed because an edit alwasy changes ident, but might not change rev
- _log
+ _log (NOT IMPLEMENTED)
history of when edits were actually applied
allows fast lookups of history of an entity (ident)
unnecessary if we keep a log of edit group accepts?
@@ -90,3 +116,37 @@ Each entity type has tables:
"wip redirect" or "wip deleted" are invalid states
+## Entity Endpoints/Actions
+
+Actions could, in theory, be directed at any of:
+
+ entities (ident)
+ revision
+ edit
+
+A design decision to be made is how much to abstract away the distinction
+between these three types (particularly the identifier/revision distinction).
+
+Top-level entity actions (resulting in edits):
+
+ create (new rev)
+ redirect
+ split
+ update (new rev)
+ delete
+
+On existing entity edits (within a group):
+
+ update
+ delete
+
+An edit group as a whole can be:
+
+ create
+ submit
+ accept
+
+Other per-entity endpoints:
+
+ match (by field/context)
+ lookup (by external persistent identifier)
diff --git a/notes/url_structure.txt b/notes/url_structure.txt
index 8c3d8ff1..14a3e41d 100644
--- a/notes/url_structure.txt
+++ b/notes/url_structure.txt
@@ -36,3 +36,4 @@ Workflow requires sessions to hold, eg, current edit group.
/ some sort of about page, html?
/v0/...
/graphql/v0/... ?
+