The term inventory database, as used within this repository, refers to a directory containing a number of JSON files:
inventory/
├── inventory.json
├── media.json
└── schema.json
These files store JSON data. They are version controlled by git. In order to allow for comprehensible diffs, they should be written in a canonical format, for instance with one item per line, 4 spaces indent, sorted keys. Such canonicalization is typically done by some post processing with a standard tool, such as jq:
jq --indent 4 < inventory.json | sponge inventory.json
An inventory database shall store properties about individual items (or records, entries). Typically, the names (keys) of these properties (values) follow some standard schema. A relational database (following the SQL industry standard, for instance PostgreSQL or SQLite) is the standard approach to model and store such data in a comprehensible manner.
Traditional database systems require a central server. Decentralized databases are subject to the CAP theorem. There is a whole industry of databases to deal with these topics. However, when I was designing a lightweight inventory managament system, I was in search for an easy replacement of a single inventory table, until then managed with Excel/OpenOffice.
Being a data scientist at heart, storing the inventory as a flat file, such as comma seperated values (CSV), was the most obvious choice for easy processing with standard scripting such as Perl, Python, Bash, etc. Storing strucutred data is just one step beyond a CSV file. In such an attempt, each CSV cell can itself store lists/dictionaries/tables/whatever-you-want. I see XML, YAML and JSON being the defacto standards for storing structured, primarily textual data in files. From all of them, JSON is especially simple and integrated in many standard libraries.
The major advantage of using a flat file database representation is the possiblity of trivially adding version control to the overall database (content/schema). By adopting git, we get decentralization and backups for free.
The major drawback of this attemp is the lack of proper foreign keys (especially accross several JSON files) and the inherent repetition (especially of keys in the JSON file). On the other hand, it is easy to check for data integrity and do conversion from/to any format, especially the obvious SQLite.
The choice of flat files is only maintainable for small datasets at the order of a few thousand entries. Such files, only a few kilobytes in size, can be parsed whenever needed in no time. At last, I choosed this format because I primarily concentrated on a web based JavaScript editor, a world in which JSON is the standard input/output file format. However, tools such as jq also bring JSON data wrangling to the command line, not to mention the vast amount of command line structured text tools.
Or, to phrase it differently: This is the ideal serialization format for inventory managament made for geeks.
The inventory.json file itself is a list of dictionaries,
where each item in the list stores one inventory item. For instance,
an example valid inventory file is:
[
{
"Bereich": "Werkstatt",
"Beschreibung": "Einfach ein Beispiel",
"Datum": "14.06.18",
"Inv-Nr.": "101",
"Label-Pos": "hinten rechts",
"Objekt": "Foo Bar Gerät",
"Ort": "M66",
"Serie": "A-3x9"
},
{
"Bereich": "Werkstatt",
"Beschreibung": "Zweiter Eintrag",
"Datum": "14.06.18",
"Inv-Nr.": "102",
"Label-Pos": "vorne unten",
"Objekt": "DEC-Schrank ABC",
"Ort": "M66",
"Serie": "A-3x9"
},
{
"Bereich": "Werkstatt",
"Beschreibung": "Mit drei Lochstreifenlesern, usw.",
"Datum": "14.06.18",
"Inv-Nr.": "103",
"Label-Pos": "vorne unten",
"Objekt": "DEC-Schrank",
"Ort": "M66",
"Serie": "A-3x9"
}
]
This file is written in German language: Both keys and values are German. Furthermore, there are a number of fields, which all are of type string in this particular example. Note the standard indent and sorting of keys.
The meaning of the keys is explained by the JSON Schema
file called schema.json. While the primary job of a Schema is to define the allowed
structure of a file, we enrich the JSON Schema with supplementary information
about the fields, such as:
examples: Exemplary field valuestitle: Human readable short description of the fieldtype: Supposed data type of the field (to distinguish, for instance,stringfrommarkdown.editor: The supposed editor widget (used for instance in the inventory editor).editor=selectindicates a drop down whereaseditor=multilineindicates a multiline string.
Here is an example of the schema file as it is used in the technikum29 inventory database:
{
"$comment": "JSON-Schema (alike) for a record in the t29 inventory table",
"$id": "http://technikum29.de/inv/schema.json",
"$schema": "http://json-schema.org/draft-04/hyper-schema",
"properties": {
"Absicht": {
"editor": "select",
"examples": [
"Teil der Ausstellung",
"Verkaufen",
"Wegwerfen/Schrott"
],
"title": "Handlungsabsicht mit diesem Inventar",
"type": "string"
},
"Bereich": {
"editor": "select",
"editor_flags": {
"restrict_options_by": "Ort"
},
"examples": [
"Keller 1",
"Werkstatt",
"Gartenhaus",
"UG"
],
"title": "Raum in der Immobilie",
"type": "string",
"order": 5
},
"Beschreibung": {
"editor": "multiline",
"examples": [
"HP85, Bildschirm defekt"
],
"title": "Kurzbeschreibung (pr\u00e4gnanter Satz), danach beliebig ausf\u00fchrlich",
"type": "markdown",
"order": 2
},
"Datum": {
"examples": [
"14.06.18"
],
"immutable": true,
"show_unique_values_in_schema_editor": true,
"title": "Datum der Erstinventarisierung",
"type": "string",
"order": 3
},
"Inhalt": {
"editor": "list",
"examples": [
[
"Buch 1",
"Buch 2",
"Buch 3"
],
[
"Netzteil XYZ",
"Bildschirm ABC",
"Prozessor DEF"
]
],
"items": {
"editor": "single-line",
"type": "string"
},
"title": "Bei mehrteiligen Inventariaten: Beschreibung der Einzelteile, zB. bei Regalen, Kisten, Ger\u00e4ten aus mehreren Bestandteilen.",
"type": "array"
},
"Inv-Nr.": {
"primary-key": true,
"title": "Inventur-Nummer: Laufende Nummer des Inventariats",
"examples": [
"102",
"H301"
],
"immutable": true,
"title": "Inventur-Nummer",
"type": "string",
"order": 0
},
"Label-Pos": {
"editor": "select",
"examples": [
"hinten rechts",
"vorne unten"
],
"title": "Position der Klebe-Inventarmarke am Ger\u00e4t",
"type": "string"
},
"Objekt": {
"counterexamples": [
"Anita 85A-Rechner",
"Dell-Combitron",
"Holzregal"
],
"editor": "select",
"editor_flags": {
"make-easy-to-add-option": true
},
"examples": [
"Tischrechner",
"Radio",
"Fernseher",
"Regal"
],
"title": "Ein Wort, welches den Objekt-Typ beschreibt",
"type": "string",
"order": 1
},
"Ort": {
"editor": "select",
"examples": [
"M66",
"AF29",
"FR4"
],
"title": "Orts-Beschreibung (nur Immobilie)",
"type": "string",
"order": 4
},
"Preis": {
"editor": "single-line",
"examples": [
"< 100\u20ac bei eBay",
"100-200EUR",
"wertlos"
],
"title": "Preisvorstellungen (gerne mit Kommentar oder Link)",
"type": "string"
},
"Zustand": {
"editor": "select",
"editor_flags": {
"make-easy-to-add-option": true
},
"examples": [
"Ohne Geh\u00e4use",
"Vergilbt",
"Einwandfrei und funktionsf\u00e4hig"
],
"title": "Knappe Beschreibung des Zustandes",
"type": "string"
}
},
"type": "object"
}