emolero commited on
Commit
ea03018
·
verified ·
1 Parent(s): b3d7a97

docs: short README (structure overview + file-map)

Browse files
Files changed (1) hide show
  1. README.md +12 -75
README.md CHANGED
@@ -71,79 +71,16 @@ configs:
71
  path: data/SAM/train.parquet
72
  ---
73
 
74
- # GSMA Specifications Corpus
75
-
76
- Permanent Reference Documents (PRDs), technical reports and white papers from the GSMA (Global System for Mobile Communications Association) working groups.
77
-
78
- Source documents (DOCX, PDF) were converted to GitHub-flavoured Markdown via the
79
- [Datalab Marker](https://www.datalab.to) API (`mode=accurate`) on 2026-05-14.
80
- Only the parsed `.md` text is published here — original binaries and extracted images are not redistributed.
81
-
82
- ## Stats
83
-
84
- - **Documents**: 253
85
- - **Working groups**: 14
86
- - **Total characters**: 62,694,319
87
- - **Total words**: 5,659,420
88
-
89
- ## Working groups
90
-
91
- Each row below corresponds to a `config_name` in the dataset viewer — pick one to browse only that
92
- working group, or pick `default` to see every document.
93
-
94
- | WG | Docs | Description |
95
- |---|---:|---|
96
- | `NG` | 95 | Networks Group — IP, IMS, roaming, voice, cloud infra (IR.* / NG.* / RCC.* PRDs) |
97
- | `TSG` | 48 | Terminals Steering Group — device testing, MIoT, TAC/IMEI allocation (TS.* PRDs) |
98
- | `FASG` | 30 | Fraud and Security Group — security baselines, NESAS scheme (FS.* PRDs) |
99
- | `eSim` | 29 | eUICC / Remote SIM Provisioning — consumer and M2M eSIM (SGP.* specs) |
100
- | `IG` | 15 | Identity Group — digital identity, Messaging-as-a-Platform, quantum identity (IG.* PRDs) |
101
- | `OPG` | 12 | Operator Platform Group — Open Gateway, Network-as-a-Service, edge APIs (OPG.* PRDs) |
102
- | `RCS` | 6 | Rich Communication Services — messaging interop, business messaging (RCC.* PRDs) |
103
- | `WAS` | 6 | Wholesale Agreements & Solutions — inter-operator wholesale settlement (WA.* / AA.* PRDs) |
104
- | `SAS` | 5 | Security Accreditation Scheme — site and UICC production audit standards |
105
- | `AID` | 2 | AI & Data — emerging AI/data PRDs (AID.*) |
106
- | `ISAG` | 2 | Interconnect Services Action Group — interconnect requirements (RCC.* PRDs) |
107
- | `IDS` | 1 | Industry Data Standards — telco data exchange documents (TD.*) |
108
- | `PQTN` | 1 | Post-Quantum Telco Network — PQC migration impact whitepapers (PQ.*) |
109
- | `SAM` | 1 | SIM Activation Management — SIM lifecycle management (SAM.*) |
110
-
111
- ## Schema
112
-
113
- | Column | Type | Description |
114
- |---|---|---|
115
- | `wg` | string | Working group code (matches the config name) |
116
- | `doc_id` | string | Document filename stem |
117
- | `title` | string | First H1 in the markdown, or `doc_id` when absent |
118
- | `rel_path` | string | Path relative to the corpus root |
119
- | `text` | string | Parsed markdown body |
120
- | `n_chars` | int64 | `len(text)` |
121
- | `n_words` | int64 | Whitespace word count of `text` |
122
- | `n_lines` | int64 | Line count of `text` |
123
-
124
- ## Loading
125
-
126
- ```python
127
- from datasets import load_dataset
128
-
129
- # Every document across every working group
130
- ds = load_dataset("GSMA/gsma", "default", split="train")
131
-
132
- # Just one working group (e.g. NG)
133
- ds_wg = load_dataset("GSMA/gsma", "NG", split="train")
134
-
135
- print(ds_wg[0]["title"])
136
- print(ds_wg[0]["text"][:500])
137
- ```
138
-
139
- ## Layout
140
-
141
- ```
142
- data/<wg>/train.parquet one parquet shard per working group
143
- README.md this dataset card
144
  ```
145
-
146
- ## License and use
147
-
148
- Private dataset; access restricted to GSMA organisation members on HuggingFace. The underlying
149
- documents are subject to GSMA / O-RAN Alliance redistribution terms — internal research use only.
 
71
  path: data/SAM/train.parquet
72
  ---
73
 
74
+ # GSMA source mirror
75
+
76
+ This dataset mirrors the GSMA's published specifications, industry guidelines, and permanent reference documents. Every source is kept twice: `original/` is the file as GSMA released it (PDF and Word), and `marked/` is that same document converted to Markdown for search and retrieval, one `raw.md` per document with any figures extracted beside it. The two trees share identical paths, so a file in `original/` has a parsed twin at the matching path under `marked/`; both are grouped by GSMA working group and programme (eSIM, Networks, RCS, Fraud and Security, and the IG/IR document series). `manifest.jsonl` carries one row per file (path, SHA-256, parse status), and `STATUS.md` reports how much of the corpus has been parsed.
77
+
78
+ ```text
79
+ gsma/
80
+ original/ # source files as published, by working group
81
+ NG/ RCS/ eSim/ IG/ FASG/ ...
82
+ marked/ # Markdown mirror of original/, same paths
83
+ NG/.../raw.md # one per document, images beside it
84
+ manifest.jsonl # one row per file: path, sha256, parse status
85
+ STATUS.md # parse-coverage summary
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
86
  ```