Add 2012-04 (129,931 items)

README.md

@@ -3,9 +3,11 @@ license: odc-by
 task_categories:
 - text-generation
 - feature-extraction
+- text-classification
+- question-answering
 language:
 - en
-pretty_name: Hacker News Open Index
+pretty_name: Hacker News — Complete Archive
 size_categories:
 - 10M<n<100M
 tags:
@@ -13,6 +15,10 @@ tags:
 - forum
 - text
 - parquet
+- community
+- tech
+- comments
+- live-updated
 configs:
 - config_name: default
   data_files:
@@ -24,85 +30,224 @@ configs:
     path: today/*.parquet
 ---
 
-# Hacker News Open Index
+# Hacker News — Complete Archive
 
-> Every Hacker News item since 2006, updated every 5 minutes — ready for training and retrieval
+Every Hacker News item since 2006, updated every 5 minutes. Stories, comments, Ask HN, Show HN, jobs, and polls — spanning **2006-10** to **2012-05** (47,332,789 items in full archive), with **4071134 items** committed so far.
 
-## What is it?
+This dataset mirrors the full [Hacker News](https://news.ycombinator.com) archive as monthly Parquet files, with 5-minute live blocks for today's activity. All fields are preserved exactly as they appear in the [HN API](https://github.com/HackerNewsAPI/HN-API).
 
-This dataset contains the full Hacker News archive: **3802518 items** spanning 1970-01 to 2026-03-14 10:12 UTC, published as monthly Parquet files with 5-minute live blocks for today.
+---
 
-Data includes stories, comments, Ask HN, Show HN, jobs, polls, and poll options — all fields preserved exactly as posted.
+## At a Glance
 
-## Dataset Stats
+| Metric | Committed | Full Archive |
+|--------|----------:|------------:|
+| Items | 4071134 | 47,332,789 |
+| Months | 67 | |
+| First month | 2006-10 | 2006-10 |
+| Last committed | 2012-05 | |
+| Size | 859.0 MB | |
+| Contributors | |  1,085,135 |
+| Stories | | 6,032,585 |
+| Comments | | 41,264,455 |
+| Last updated | 2026-03-14 10:21 UTC | |
+
+---
+
+## Growth Over Time
+
+Items committed to this dataset by year:
+
+```
+  2006  █░░░░░░░░░░░░░░░░░░░░░░░░░░░░░  62
+  2007  ██░░░░░░░░░░░░░░░░░░░░░░░░░░░░  93.8K
+  2008  ███████░░░░░░░░░░░░░░░░░░░░░░░  320.9K
+  2009  █████████████░░░░░░░░░░░░░░░░░  608.4K
+  2010  ██████████████████████░░░░░░░░  1.0M
+  2011  ██████████████████████████████  1.4M
+  2012  ██████████████░░░░░░░░░░░░░░░░  637.5K
+```
+
+---
+
+## Content Breakdown
+
+| Type | Count | Share |
+|------|------:|------:|
+| comment | 41,264,455 | 87.2% |
+| story | 6,032,585 | 12.7% |
+| job | 18,065 | 0.0% |
+| poll | 2,239 | 0.0% |
+| pollopt | 15,445 | 0.0% |
+
+**84.8%** of stories link to external URLs. The remaining are text posts (Ask HN, Show HN, Launch HN).
+
+Average discussion depth: **23.9 comments** per story (max: 9,275).
+
+---
+
+## Story Scores
 
 | Metric | Value |
-|--------|-------|
-| Total items | 3802518 |
-| Historical months | 65 |
-| First month | 1970-01 |
-| Last committed month | 2012-03 |
-| Total size | 793.5 MB |
-| Last updated | 2026-03-14 10:12 UTC |
+|--------|------:|
+| Average score | 1.5 |
+| Median score | 0 |
+| Highest score ever | 6,015 |
+| Stories with 100+ points | 175,896 |
+| Stories with 1,000+ points | 2,169 |
+
+Most stories receive a handful of points, but the distribution has a long tail. The top 0.03% of stories (1,000+ points) represent the most discussed content in tech.
+
+---
+
+## Most-Shared Domains
+
+The most frequently linked domains across all stories:
 
-## File Layout
+| # | Domain | Stories |
+|--:|--------|--------:|
+| 1 | github.com | 196,963 |
+| 2 | www.youtube.com | 134,704 |
+| 3 | medium.com | 124,515 |
+| 4 | www.nytimes.com | 77,608 |
+| 5 | en.wikipedia.org | 54,362 |
+| 6 | techcrunch.com | 54,178 |
+| 7 | twitter.com | 50,477 |
+| 8 | arstechnica.com | 47,040 |
+| 9 | www.theguardian.com | 44,227 |
+| 10 | www.bloomberg.com | 37,766 |
 
+---
+
+## Most Active Story Submitters
+
+Users ranked by total story submissions:
+
+| # | User | Stories |
+|--:|------|--------:|
+| 1 | rbanffy | 36,762 |
+| 2 | Tomte | 26,154 |
+| 3 | tosh | 24,003 |
+| 4 | bookofjoe | 20,548 |
+| 5 | mooreds | 20,330 |
+| 6 | pseudolus | 19,901 |
+| 7 | PaulHoule | 18,930 |
+| 8 | todsacerdoti | 18,880 |
+| 9 | ingve | 17,042 |
+| 10 | thunderbong | 15,945 |
+| 11 | jonbaer | 14,161 |
+| 12 | rntn | 13,410 |
+| 13 | doener | 12,779 |
+| 14 | Brajeshwar | 12,274 |
+| 15 | LinuxBender | 11,058 |
+
+---
+
+## Quick Start
+
+### DuckDB
+
+The fastest way to explore the dataset. DuckDB reads Parquet natively from Hugging Face:
+
+```sql
+-- Top 20 highest-scored stories of all time
+SELECT id, title, by, score, url, time
+FROM read_parquet('hf://datasets/open-index/hacker-news/data/*/*.parquet')
+WHERE type = 'story' AND title != ''
+ORDER BY score DESC
+LIMIT 20;
+```
+
+```sql
+-- Monthly submission volume for 2024
+SELECT
+    toStartOfMonth(time) AS month,
+    count(*) AS items,
+    countIf(type = 'story') AS stories,
+    countIf(type = 'comment') AS comments
+FROM read_parquet('hf://datasets/open-index/hacker-news/data/2024/*.parquet')
+GROUP BY month
+ORDER BY month;
 ```
-data/
-  2006/2006-10.parquet    ← first HN month
-  ...
-  2026/2026-02.parquet
-today/
-  2026-03-14_00_00.parquet   ← 5-min live blocks
-  2026-03-14_00_05.parquet
-  ...
-stats.csv                    ← one row per committed month
-stats_today.csv              ← one row per committed 5-min block
+
+```sql
+-- Most discussed stories (by comment count) in the last year
+SELECT id, title, by, score, descendants AS comments, url, time
+FROM read_parquet('hf://datasets/open-index/hacker-news/data/2025/*.parquet')
+WHERE type = 'story' AND descendants > 0
+ORDER BY descendants DESC
+LIMIT 20;
+```
+
+```sql
+-- Who posts the most Ask HN questions?
+SELECT by, count(*) AS posts
+FROM read_parquet('hf://datasets/open-index/hacker-news/data/*/*.parquet')
+WHERE type = 'story' AND title LIKE 'Ask HN:%'
+GROUP BY by
+ORDER BY posts DESC
+LIMIT 20;
 ```
 
-## How to Use
+```sql
+-- Domain popularity over time
+SELECT
+    extract(year FROM time) AS year,
+    regexp_extract(url, 'https?://([^/]+)', 1) AS domain,
+    count(*) AS stories
+FROM read_parquet('hf://datasets/open-index/hacker-news/data/*/*.parquet')
+WHERE type = 'story' AND url != ''
+GROUP BY year, domain
+QUALIFY row_number() OVER (PARTITION BY year ORDER BY stories DESC) <= 5
+ORDER BY year, stories DESC;
+```
 
 ### Python (datasets)
 
 ```python
 from datasets import load_dataset
 
-# Stream the full history
+# Stream the full history (no download required)
 ds = load_dataset("open-index/hacker-news", split="train", streaming=True)
 for item in ds:
     print(item["id"], item["type"], item["title"])
 
-# Load a single month
+# Load a specific year into memory
 ds = load_dataset(
     "open-index/hacker-news",
-    data_files="data/2024/2024-01.parquet",
+    data_files="data/2024/*.parquet",
     split="train",
 )
+print(f"{len(ds):,} items in 2024")
 ```
 
-### DuckDB
+### Python (pandas + DuckDB)
 
-```sql
--- All stories from 2024
-SELECT id, by, title, score, url
-FROM read_parquet('hf://datasets/open-index/hacker-news/data/2024/*.parquet')
-WHERE type = 'story'
-ORDER BY score DESC
-LIMIT 20;
+```python
+import duckdb
+
+conn = duckdb.connect()
 
--- Live blocks for today
-SELECT id, by, title, time
-FROM read_parquet('hf://datasets/open-index/hacker-news/today/*.parquet')
-ORDER BY id DESC
-LIMIT 50;
+# Compute score percentiles
+df = conn.sql("""
+    SELECT
+        percentile_disc(0.50) WITHIN GROUP (ORDER BY score) AS p50,
+        percentile_disc(0.90) WITHIN GROUP (ORDER BY score) AS p90,
+        percentile_disc(0.99) WITHIN GROUP (ORDER BY score) AS p99,
+        percentile_disc(0.999) WITHIN GROUP (ORDER BY score) AS p999
+    FROM read_parquet('hf://datasets/open-index/hacker-news/data/*/*.parquet')
+    WHERE type = 'story'
+""").df()
+print(df)
 ```
 
-### huggingface_hub
+### Download Specific Files
 
 ```python
 from huggingface_hub import snapshot_download
 
-folder = snapshot_download(
+# Download only 2024 data
+snapshot_download(
     "open-index/hacker-news",
     repo_type="dataset",
     local_dir="./hn/",
@@ -110,28 +255,106 @@ folder = snapshot_download(
 )
 ```
 
+```bash
+# CLI: download a single month
+huggingface-cli download open-index/hacker-news \
+    data/2024/2024-01.parquet \
+    --repo-type dataset --local-dir ./hn/
+```
+
+---
+
 ## Schema
 
+Every Parquet file shares the same schema, matching the [HN API](https://github.com/HackerNewsAPI/HN-API) item format:
+
 | Column | Type | Description |
 |--------|------|-------------|
-| `id` | int64 | Item ID (monotonically increasing) |
-| `deleted` | bool | Soft-deleted flag |
-| `type` | string | story, comment, ask, show, job, poll, pollopt |
-| `by` | string | Username of author |
-| `time` | DateTime | Post timestamp (UTC) |
-| `text` | string | HTML body (comments, Ask HN, jobs) |
-| `dead` | bool | Flagged/killed by moderators |
-| `parent` | int64 | Parent item ID (for comments) |
-| `poll` | int64 | Poll item ID (for pollopts) |
-| `kids` | Array(int64) | Direct child item IDs |
-| `url` | string | External URL (stories) |
-| `score` | int64 | Points |
-| `title` | string | Story/Ask/Show/Job title |
-| `parts` | Array(int64) | Poll option IDs |
-| `descendants` | int64 | Total descendant comment count |
+| `id` | int64 | Unique item ID (monotonically increasing) |
+| `deleted` | bool | `true` if soft-deleted by author or moderators |
+| `type` | string | One of: `story`, `comment`, `job`, `poll`, `pollopt` |
+| `by` | string | Username of the author |
+| `time` | timestamp | Post timestamp (UTC) |
+| `text` | string | HTML body text (comments, Ask HN, jobs) |
+| `dead` | bool | `true` if flagged/killed by moderators |
+| `parent` | int64 | Parent item ID (comments only) |
+| `poll` | int64 | Associated poll ID (poll options only) |
+| `kids` | list\<int64\> | Direct child item IDs |
+| `url` | string | External URL (stories with links) |
+| `score` | int64 | Points (upvotes minus downvotes) |
+| `title` | string | Title text (stories, jobs, polls) |
+| `parts` | list\<int64\> | Poll option item IDs (polls only) |
+| `descendants` | int64 | Total comment count in the discussion tree |
+
+---
+
+## Data Architecture
+
+### File Layout
+
+```
+open-index/hacker-news/
+  data/
+    2006/2006-10.parquet      # first HN month with data
+    2006/2006-12.parquet
+    2007/2007-01.parquet
+    ...
+    2026/2026-02.parquet      # most recent complete month
+  today/
+    2026-03-14_00_00.parquet  # 5-minute live blocks
+    2026-03-14_00_05.parquet
+    ...
+  stats.csv                   # one row per committed month
+  stats_today.csv             # one row per committed 5-min block
+  README.md
+```
+
+### Update Pipeline
+
+This dataset is maintained by an automated pipeline:
+
+1. **Historical backfill** — On first run, fetches every month from 2006-10 to the most recent complete month. Each month is committed as a single Parquet file. Months already committed (tracked in `stats.csv`) are skipped, making the process resumable.
+
+2. **Live polling** — After backfill, polls the source every 5 minutes for new items. Each batch is committed as `today/YYYY-MM-DD_HH_MM.parquet`.
+
+3. **Day rollover** — At midnight UTC, all of today's 5-minute blocks are merged into the monthly Parquet file using DuckDB, then the individual blocks are deleted.
+
+Data is sourced from the [ClickHouse Playground](https://sql.clickhouse.com) which mirrors the official HN Firebase API.
+
+### Parquet Compression
+
+All files use **Zstandard (zstd)** compression at level 22 for optimal size. Monthly files are sorted by `id` for efficient range scans.
+
+---
+
+## Dataset Stats
+
+Detailed per-month statistics are available in `stats.csv`:
+
+| Column | Description |
+|--------|-------------|
+| `year`, `month` | Calendar month |
+| `lowest_id`, `highest_id` | Item ID range in this file |
+| `count` | Number of items |
+| `dur_fetch_s` | Seconds to fetch from source |
+| `dur_commit_s` | Seconds to commit to Hugging Face |
+| `size_bytes` | Parquet file size |
+| `committed_at` | ISO 8601 commit timestamp |
+
+```sql
+-- Query the stats directly
+SELECT * FROM read_csv_auto('hf://datasets/open-index/hacker-news/stats.csv')
+ORDER BY year, month;
+```
+
+---
 
 ## License
 
 Released under the **Open Data Commons Attribution License (ODC-By) v1.0**.
-Original content is subject to the rights of its respective authors.
-Hacker News data is provided by Y Combinator.
+
+Original content is subject to the rights of its respective authors. Hacker News data is provided by [Y Combinator](https://www.ycombinator.com). This dataset is an independent mirror — it is not affiliated with or endorsed by Y Combinator.
+
+---
+
+*Last updated: 2026-03-14 10:21 UTC*

data/2012/2012-04.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aecddff0f7ab68068430894a8d0e46b1956a4dcd780a3a763e96a8a368b934f9
+size 33171230

stats.csv

@@ -63,4 +63,6 @@ year,month,lowest_id,highest_id,count,dur_fetch_s,dur_commit_s,size_bytes,commit
 2011,12,3297497,3412253,114578,2,7,25620474,2026-03-14T10:12:20Z
 2012,1,3412254,3535903,123265,3,15,32611036,2026-03-14T10:12:30Z
 2012,2,3535904,3650342,114197,3,6,25435313,2026-03-14T10:12:49Z
-2012,3,3650343,3782006,131425,2,0,33789898,2026-03-14T10:12:57Z
+2012,3,3650343,3782006,131425,2,12,33789898,2026-03-14T10:12:57Z
+2012,4,3782007,3912227,129931,3,0,33171230,2026-03-14T10:20:32Z
+2012,5,3912228,4051124,138685,8,0,35491708,2026-03-14T10:21:15Z