Add detailed dataset card and the scraping pipeline used to build it

README.md

@@ -18,6 +18,8 @@ tags:
 - news
 - news-articles
 - opinion-journalism
+- research-only
+- text-and-data-mining
 - web-scraping
 - weekly-newspaper
 annotations_creators:
@@ -68,8 +70,17 @@ a set of comparable corpora covering the German media landscape from public-serv
 broadcasters to alternative and fringe outlets, all built with the same crawler and the same
 extraction pipeline so that they can be compared and combined directly.
 
-> ⚠️ **Research material.** The articles are reproduced verbatim from the publisher's website.
-> Their content expresses the views of that publisher, not of the dataset author.
+> ## ⚠️ For scientific research only
+>
+> This corpus is published **exclusively for scientific research** under the German
+> research exception for text and data mining — the *Forschungsprivileg* of
+> **§ 60d UrhG** (implementing Art. 3 of EU Directive 2019/790). The article texts remain
+> the copyright of their publisher and are **not** licensed for commercial use,
+> redistribution, or republication. By downloading it you confirm that you use it for
+> non-commercial scientific research. See [Licensing and intended use](#licensing-and-intended-use).
+>
+> The articles are reproduced verbatim from the publisher's website; their content expresses
+> the views of that publisher, not of the dataset author.
 
 ## At a glance
 
@@ -82,6 +93,7 @@ extraction pipeline so that they can be compared and combined directly.
 | **Split** | `train` (single split) |
 | **Format** | Parquet, 1 shard, ≈ 89 MB download / 138 MB in memory |
 | **Perceived partisanship** | 2.7 / 7 → *left of centre* ([medienkompass.org](https://medienkompass.org/deutsche-medienlandschaft/) reader survey) |
+| **Intended use** | Scientific research only (*Forschungsprivileg*, § 60d UrhG) |
 | **Mirror** | [`NLP-UniBW/der_freitag_de`](https://huggingface.co/datasets/NLP-UniBW/der_freitag_de) |
 
 ```python
@@ -440,13 +452,44 @@ corpus = concatenate_datasets(parts)
 
 ## Licensing and intended use
 
-The crawling and extraction code in `scraping/` is released for reuse. The **article texts remain
-the copyright of der Freitag / its publisher**; they are reproduced here for non-commercial
-research purposes such as text mining, media-bias research, discourse analysis and language-model
-evaluation. If you intend to use the texts in another setting — redistribution, commercial use,
-or publication of substantial verbatim excerpts — check German copyright law (in particular
-§ 60d UrhG on text and data mining for research) and the publisher's terms of use. If you are a
-rights holder and want content removed, please open a discussion on this repository.
+### Scientific research only — *Forschungsprivileg* (§ 60d UrhG)
+
+This dataset is made available **solely for the purpose of scientific research**. The legal basis
+is the German text-and-data-mining exception for research, the so-called **Forschungsprivileg**:
+
+- **§ 60d UrhG** — *Text und Data Mining für Zwecke der wissenschaftlichen Forschung* — allows
+  research organisations to make reproductions of lawfully accessible works in order to carry out
+  text and data mining for **non-commercial** scientific research, and to share the resulting
+  corpus with a specifically limited circle of people for joint research and with individual third
+  parties for the purpose of quality assurance.
+- It implements **Art. 3 of Directive (EU) 2019/790** (the DSM Directive); comparable research
+  exceptions exist in the other EU member states.
+- § 60d UrhG does **not** cover commercial exploitation. Research organisations that give a
+  commercial enterprise preferential access are explicitly excluded from the privilege.
+
+### What this means for you
+
+✅ **Permitted** — using the corpus for non-commercial scientific research: text and data mining,
+corpus linguistics, media-bias and discourse analysis, training and evaluating research models,
+reproducing published findings, and quality checks by individual third parties.
+
+❌ **Not permitted** — commercial use or commercial model training; redistributing or republishing
+the article texts; presenting the texts as your own; building products or services on the article
+texts; any use that substitutes for reading the publisher's own website.
+
+The **article texts remain the copyright of der Freitag and its publisher.** No licence to them is
+granted or implied by this repository. Only the crawling and extraction code in `scraping/` is
+offered for free reuse.
+
+If your intended use falls outside the research exception, obtain permission from the rights
+holder, or use only the derived, non-expressive data (statistics, embeddings, counts). Note that
+§ 60d UrhG also expects research corpora to be deleted once they are no longer needed for the
+research purpose, with archiving permitted at trusted institutions.
+
+If you are a rights holder and want content removed, please open a discussion on this repository.
+
+> This section is an orientation for researchers, not legal advice. Responsibility for your own
+> use of the data remains with you and your institution.
 
 ## Citation
 

scraping/.env.example

@@ -0,0 +1,52 @@
+# ---------------------------------------------------------------------------
+#  crawler.py configuration for www.freitag.de
+#
+#  Copy this file to `.env` and uncomment the lines you need:
+#
+#      cp .env.example .env
+#
+#  Variables that are already set in the real environment take precedence over
+#  this file, so a single run can still be overridden on the command line:
+#
+#      SCRAPER_THREADS=4 python crawler.py
+#
+#  Everything here is optional - the defaults shown are exactly the values that
+#  were used for the original crawl.
+# ---------------------------------------------------------------------------
+
+
+# === Proxies ===============================================================
+
+# Inline pool. Separate entries with commas, semicolons, spaces or newlines.
+# A scheme is optional - "http://" is assumed when none is given. One entry is
+# picked per request, so several entries spread the load over several exits.
+#SCRAPER_PROXIES=proxy-a.example.com:3128,proxy-b.example.com:3128
+
+# ... with an explicit scheme and credentials:
+#SCRAPER_PROXIES=http://user:pass@proxy.example.com:3128
+
+# ... SOCKS (requires PySocks):
+#SCRAPER_PROXIES=socks5://localhost:9050
+
+# For a whole proxy farm, keep the list in its own file instead: one proxy per
+# line, blank lines and "#" comments allowed. See `proxies.example.txt`.
+# This scales to thousands of entries and can be regenerated by another tool
+# without ever touching this file. Both settings may be combined.
+#SCRAPER_PROXY_FILE=proxies.txt
+
+# Leave both unset to crawl over a direct connection - that is the default.
+
+
+# === Crawler tuning ========================================================
+
+# Number of worker processes sharing the URL frontier.
+#SCRAPER_THREADS=32
+
+# Socket timeout per request, in seconds.
+#SCRAPER_TIMEOUT=10.0
+
+# Attempts per URL before it is given up on.
+#SCRAPER_RETRIES=10
+
+# Seed URL the crawl starts from. Any article URL on the site works.
+#SCRAPER_START_URL=https://www.freitag.de/autoren/franz-schandl/asylpolitik-in-oesterreich-drangsalieren-mehr-faellt-ihnen-nicht-ein

scraping/README.md

@@ -4,6 +4,8 @@
 |---|---|
 | [`crawler.py`](./crawler.py) | Distributed whole-site crawler for `www.freitag.de`. Writes one JSON file per article to `pages/`. |
 | [`finalize.ipynb`](./finalize.ipynb) | Merges `pages/*.json` into a single table, deduplicates it and pushes it to the Hub. |
+| [`.env.example`](./.env.example) | Configuration template. Copy to `.env` and uncomment what you need. |
+| [`proxies.example.txt`](./proxies.example.txt) | Proxy-list template for `SCRAPER_PROXY_FILE` — one proxy per line. |
 
 ```bash
 pip install mechanicalsoup beautifulsoup4 lxml python-dateutil fake-useragent \
@@ -13,19 +15,44 @@ python crawler.py          # writes pages/*.json; `touch killall.txt` to stop
 jupyter notebook finalize.ipynb
 ```
 
-## Using your own proxies
+## Configuration
+
+All settings are optional and are read from the environment or from a `.env`
+file next to `crawler.py`. The defaults are exactly the values used for the
+original crawl.
+
+```bash
+cp .env.example .env     # then uncomment what you need
+```
+
+| Variable | Meaning |
+|---|---|
+| `SCRAPER_PROXIES` | Inline proxy pool — entries separated by commas, semicolons, spaces or newlines. |
+| `SCRAPER_PROXY_FILE` | Path to a file with **one proxy per line** — the practical form for a proxy farm. |
+| `SCRAPER_THREADS` | Number of worker processes sharing the URL frontier. |
+| `SCRAPER_TIMEOUT` | Socket timeout per request, in seconds. |
+| `SCRAPER_RETRIES` | Attempts per URL before giving up. |
+| `SCRAPER_START_URL` | Seed URL the crawl starts from. |
+
+Environment variables take precedence over `.env`, so a single run can be
+overridden on the command line:
+
+```bash
+SCRAPER_THREADS=4 python crawler.py
+```
+
+### Proxies
 
 The original crawl ran behind a pool of forward proxies on a private network.
-Those addresses have been removed from this published copy. `crawler.py` reads
-the `SCRAPER_PROXIES` environment variable instead — a comma-separated list of
-proxies, with or without a scheme (`http://` is assumed when none is given).
-One entry is picked per request, so a pool spreads the load over several exits.
+Those addresses have been removed from this published copy. A scheme is
+optional on every entry (`http://` is assumed); one entry is picked per
+request, so a pool spreads the load over several exits.
 
 ```bash
 # a single proxy
 export SCRAPER_PROXIES="proxy-a.example.com:3128"
 
-# a pool, rotated per request
+# a small pool, rotated per request
 export SCRAPER_PROXIES="proxy-a.example.com:3128,proxy-b.example.com:3128"
 
 # with an explicit scheme and credentials
@@ -33,13 +60,22 @@ export SCRAPER_PROXIES="http://user:pass@proxy.example.com:3128"
 
 # SOCKS, if PySocks is installed
 export SCRAPER_PROXIES="socks5://localhost:9050"
+```
 
-# back to a direct connection (the default)
-unset SCRAPER_PROXIES
+For a **whole proxy farm**, keep the list in its own file — one proxy per line,
+blank lines and `#` comments allowed — and point `SCRAPER_PROXY_FILE` at it.
+The file can hold thousands of entries and be regenerated by another tool
+without touching `.env` or the script:
 
+```bash
+cp proxies.example.txt proxies.txt   # then fill in your own entries
+echo 'SCRAPER_PROXY_FILE=proxies.txt' >> .env
 python crawler.py
 ```
 
+Leave both unset to crawl over a direct connection — that is the default and
+what most of the crawls in this collection used.
+
 The scripts contain no credentials, cookies or personal data.
 
 Please crawl responsibly: honour `robots.txt` and the site's terms of use, keep

scraping/crawler.py

@@ -23,9 +23,14 @@ Usage
     pip install mechanicalsoup beautifulsoup4 lxml python-dateutil fake-useragent
     python crawler.py
 
-Set ``SCRAPER_PROXIES`` (comma separated ``host:port`` list) to route requests
-through your own forward proxies; see the note further down. The private proxy
-addresses used in the original crawl have been removed from this published copy.
+Configuration
+-------------
+Copy ``.env.example`` to ``.env`` and uncomment what you need - proxies, worker
+count, timeout, retries and the seed URL are all read from there (environment
+variables still win over the file). For a large proxy farm put one proxy per
+line into a text file and point ``SCRAPER_PROXY_FILE`` at it; see
+``proxies.example.txt``. The private proxy addresses used in the original crawl
+have been removed from this published copy.
 
 Please crawl responsibly: honour robots.txt and the site's terms of use, and
 lower the worker count unless you have the operator's permission.
@@ -46,49 +51,81 @@ from dateutil.parser import parse
 
 base_url = "https://www.freitag.de/"
 
-# --- Proxy configuration ---------------------------------------------------
-# The original crawl was executed behind a pool of forward proxies located on a
-# private network.  Those addresses are infrastructure specific and have been
-# removed from this published copy of the script.
+# --- Configuration ---------------------------------------------------------
+# Every setting below can be supplied through the environment or through a
+# ".env" file placed next to this script.  Copy ".env.example" to ".env" and
+# uncomment what you need:
 #
-# Set the environment variable SCRAPER_PROXIES to a comma separated list of
-# proxies to send the crawl through your own pool.  Entries may be written with
-# or without a scheme ("http://" is assumed when none is given).  Each request
-# picks one entry from the pool, so several entries spread the load over
-# several exits.  Leave the variable unset or empty to connect directly - that
-# is the default and what most of the crawls in this collection used.
+#     cp .env.example .env
 #
-# Examples:
-#
-#     # a single proxy
-#     export SCRAPER_PROXIES="proxy-a.example.com:3128"
-#
-#     # a pool, rotated per request
-#     export SCRAPER_PROXIES="proxy-a.example.com:3128,proxy-b.example.com:3128"
-#
-#     # with an explicit scheme and credentials
-#     export SCRAPER_PROXIES="http://user:pass@proxy.example.com:3128"
-#
-#     # SOCKS also works if PySocks is installed
-#     export SCRAPER_PROXIES="socks5://localhost:9050"
-#
-#     # back to a direct connection
-#     unset SCRAPER_PROXIES
-#
-#     python crawler.py
+# Nothing has to be set - the defaults are exactly the values that were used
+# for the original crawl.
+def _load_dotenv(path=".env"):
+    """Read a ".env" file: KEY=VALUE per line, "#" starts a comment.
+
+    Variables that are already set in the real environment win, so a single run
+    can still be overridden with e.g. ``SCRAPER_THREADS=4 python crawler.py``.
+    """
+    if not os.path.isfile(path):
+        return
+    with io.open(path, "r", encoding="utf-8") as handle:
+        for line in handle:
+            line = line.strip()
+            if not line or line.startswith("#") or "=" not in line:
+                continue
+            key, _, value = line.partition("=")
+            key = key.strip()
+            value = value.strip().strip('"').strip("'")
+            if key and key not in os.environ:
+                os.environ[key] = value
+
+
+_load_dotenv()
+
+
 def _load_proxy_pool():
-    """Read the SCRAPER_PROXIES environment variable into a list of proxy URLs."""
+    """Build the proxy pool from SCRAPER_PROXIES and/or SCRAPER_PROXY_FILE.
+
+    SCRAPER_PROXIES     inline list; entries separated by commas, semicolons,
+                        spaces or newlines.
+    SCRAPER_PROXY_FILE  path to a file holding one proxy per line, with blank
+                        lines and "#" comments allowed.  This is the practical
+                        form for a whole proxy farm: the pool can be swapped or
+                        regenerated without touching the script or the .env.
+
+    Both may be combined.  Entries may be written with or without a scheme
+    ("http://" is assumed when none is given).  Duplicates are dropped and an
+    empty pool simply means "connect directly".
+    """
+    raw = os.environ.get("SCRAPER_PROXIES", "")
+    proxy_file = os.environ.get("SCRAPER_PROXY_FILE", "").strip()
+    if proxy_file:
+        if not os.path.isfile(proxy_file):
+            raise SystemExit("SCRAPER_PROXY_FILE points to a missing file: " + proxy_file)
+        with io.open(proxy_file, "r", encoding="utf-8") as handle:
+            raw = raw + "\n" + handle.read()
+    cleaned = []
+    for line in raw.splitlines():
+        line = line.split("#", 1)[0].strip()
+        if line:
+            cleaned.append(line)
     pool = []
-    for entry in os.environ.get("SCRAPER_PROXIES", "").split(","):
+    for entry in re.split(r"[,;\s]+", " ".join(cleaned)):
         entry = entry.strip()
         if not entry:
             continue
         if "://" not in entry:
             entry = "http://" + entry
-        pool.append(entry)
+        if entry not in pool:
+            pool.append(entry)
     return pool
 
+
 PROXY_POOL = _load_proxy_pool()
+NUM_THREADS = int(os.environ.get("SCRAPER_THREADS", "32"))
+REQUEST_TIMEOUT = float(os.environ.get("SCRAPER_TIMEOUT", "10.0"))
+MAX_RETRIES = int(os.environ.get("SCRAPER_RETRIES", "10"))
+START_URL = os.environ.get("SCRAPER_START_URL", "https://www.freitag.de/autoren/franz-schandl/asylpolitik-in-oesterreich-drangsalieren-mehr-faellt-ihnen-nicht-ein")
 
 def log(number, url):
     if not os.path.isdir("log"):
@@ -247,7 +284,7 @@ def crawler(iId, iNumberOfThreads, lQueues):
             (not ".epub" in sPage):
 
             log(iId, sPage)
-            for i in range(10):
+            for i in range(MAX_RETRIES):
                 try:
                     lproxies = PROXY_POOL
                     proxy = lproxies[int((int(time.time())/10))%len(lproxies)] if lproxies else None
@@ -255,7 +292,7 @@ def crawler(iId, iNumberOfThreads, lQueues):
                     browser = mechanicalsoup.StatefulBrowser()
                     browser.session.proxies = proxies
                     
-                    browser.open(sPage, timeout=10.0)
+                    browser.open(sPage, timeout=REQUEST_TIMEOUT)
                     page = browser.get_current_page()
                     page_str = str(page)
                     if "#challenge-error-text" in page_str:
@@ -300,7 +337,7 @@ def main():
     if not os.path.isdir("pages"):
         os.mkdir("pages")
 
-    iNumberOfThreads = 32
+    iNumberOfThreads = NUM_THREADS
     lQueues = []
     mgr = multiprocessing.Manager()
     for i in range(iNumberOfThreads):
@@ -310,7 +347,7 @@ def main():
     lIds = range(0, iNumberOfThreads)
     lNumberOfThreads = [iNumberOfThreads] * iNumberOfThreads
     llQueues = [lQueues] * iNumberOfThreads
-    addStartChannel("https://www.freitag.de/autoren/franz-schandl/asylpolitik-in-oesterreich-drangsalieren-mehr-faellt-ihnen-nicht-ein", iNumberOfThreads, lQueues)
+    addStartChannel(START_URL, iNumberOfThreads, lQueues)
     lArguments = list(zip(lIds, lNumberOfThreads, llQueues))
     with multiprocessing.Pool(processes=iNumberOfThreads) as pool:
         pool.starmap(crawler, lArguments)

scraping/proxies.example.txt

@@ -0,0 +1,15 @@
+# One proxy per line. Blank lines and "#" comments are ignored.
+# A scheme is optional - "http://" is assumed when none is given.
+#
+# Copy this file (e.g. to `proxies.txt`), fill in your own entries and point
+# SCRAPER_PROXY_FILE at it:
+#
+#     SCRAPER_PROXY_FILE=proxies.txt
+#
+# This is the convenient form for a large proxy farm - the file can hold
+# thousands of lines and be regenerated by another tool at any time.
+
+proxy-a.example.com:3128
+proxy-b.example.com:3128
+http://user:pass@proxy-c.example.com:3128
+socks5://localhost:9050