add comments/2005-12 2005/12 (1 shards, 1.1K rows)

README.md

@@ -0,0 +1,105 @@
+---
+configs:
+- config_name: comments
+  data_files:
+  - split: train
+    path: "data/comments/**/*.parquet"
+- config_name: submissions
+  data_files:
+  - split: train
+    path: "data/submissions/**/*.parquet"
+license: other
+language:
+- en
+tags:
+- reddit
+- social-media
+- arctic-shift
+- pushshift
+pretty_name: Arctic Shift Reddit Archive
+size_categories:
+- 100B<n<1T
+---
+
+# Arctic Shift Reddit Archive
+
+Full Reddit dataset (comments + submissions) sourced from the
+[Arctic Shift](https://github.com/ArthurHeitmann/arctic_shift) project,
+covering all subreddits from 2005-06 through **2005-12**.
+
+Data is organized as monthly parquet shards by type, making it easy to load
+specific time ranges or work with comments and submissions independently.
+
+## Quick Start
+
+```python
+from datasets import load_dataset
+
+# Stream all comments (recommended — dataset is very large)
+comments = load_dataset("open-index/arctic", "comments", streaming=True)
+for item in comments["train"]:
+    print(item["author"], item["body"][:80])
+
+# Load submissions for a specific year
+subs = load_dataset("open-index/arctic", "submissions",
+                    data_files="data/submissions/2020/**/*.parquet")
+```
+
+## Dataset Stats
+
+| Type        | Months | Rows | Parquet Size |
+|-------------|--------|------|--------------|
+| comments    | 1 | 1.1K | 141717 B |
+| submissions | 0 | 0 | 0 B |
+
+*Updated: 2026-03-14*
+
+## Growth (rows per year, comments + submissions combined)
+
+```
+2005  ████████████████████████████████████████  1.1K
+```
+
+## Schema
+
+### Comments
+
+| Column | Type | Description |
+|--------|------|-------------|
+| id | VARCHAR | Comment ID |
+| author | VARCHAR | Username |
+| subreddit | VARCHAR | Subreddit name |
+| body | VARCHAR | Comment text |
+| score | BIGINT | Net upvotes |
+| created_utc | BIGINT | Unix timestamp |
+| created_at | TIMESTAMP | Derived from created_utc |
+| body_length | BIGINT | Character count of body |
+| link_id | VARCHAR | Parent submission ID |
+| parent_id | VARCHAR | Parent comment or submission ID |
+| distinguished | VARCHAR | mod/admin/null |
+| author_flair_text | VARCHAR | Author flair |
+
+### Submissions
+
+| Column | Type | Description |
+|--------|------|-------------|
+| id | VARCHAR | Submission ID |
+| author | VARCHAR | Username |
+| subreddit | VARCHAR | Subreddit name |
+| title | VARCHAR | Post title |
+| selftext | VARCHAR | Post body (self posts) |
+| score | BIGINT | Net upvotes |
+| created_utc | BIGINT | Unix timestamp |
+| created_at | TIMESTAMP | Derived from created_utc |
+| title_length | BIGINT | Character count of title |
+| num_comments | BIGINT | Comment count |
+| url | VARCHAR | External URL or permalink |
+| over_18 | BOOLEAN | NSFW flag |
+| link_flair_text | VARCHAR | Post flair |
+| author_flair_text | VARCHAR | Author flair |
+
+## Source & License
+
+Repackaged from [Arctic Shift](https://github.com/ArthurHeitmann/arctic_shift) monthly dumps,
+which re-process the [PushShift](https://pushshift.io) Reddit archive.
+Original content by Reddit users.

data/comments/2005/12/000.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4124dfc42fbccfbcc158693151f122d8af026a9c401f743e551fbda55a5b61de
+size 141717

stats.csv

@@ -0,0 +1,2 @@
+year,month,type,shards,count,size_bytes,dur_download_s,dur_process_s,dur_commit_s,committed_at
+2005,12,comments,1,1075,141717,36.67,1.04,0.00,2026-03-14T19:45:54Z