From 50f81581a3d50a2a6a6fb0e55fbdf178897b2818 Mon Sep 17 00:00:00 2001
From: Bryan Newbold <bnewbold@archive.org>
Date: Wed, 1 Apr 2020 15:30:05 -0700
Subject: start python module directory

---
 fatcat_covid19/__init__.py |  0
 fatcat_covid19/common.py   | 88 ++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 88 insertions(+)
 create mode 100644 fatcat_covid19/__init__.py
 create mode 100644 fatcat_covid19/common.py

(limited to 'fatcat_covid19')

diff --git a/fatcat_covid19/__init__.py b/fatcat_covid19/__init__.py
new file mode 100644
index 0000000..e69de29
diff --git a/fatcat_covid19/common.py b/fatcat_covid19/common.py
new file mode 100644
index 0000000..97bc675
--- /dev/null
+++ b/fatcat_covid19/common.py
@@ -0,0 +1,88 @@
+
+"""
+Helper routines.
+
+Many of these copied verbatim from fatcat or sandcrawler repositories.
+"""
+
+import os
+import sys
+import copy
+import json
+import magic
+import hashlib
+
+import requests
+from requests.adapters import HTTPAdapter
+from requests.packages.urllib3.util.retry import Retry # pylint: disable=import-error
+
+
+def gen_file_metadata(blob):
+    """
+    Takes a file blob (bytestream) and returns hashes and other metadata.
+
+    Returns a dict: size_bytes, md5hex, sha1hex, sha256hex, mimetype
+    """
+    assert blob
+    mimetype = magic.Magic(mime=True).from_buffer(blob)
+    hashes = [
+        hashlib.sha1(),
+        hashlib.sha256(),
+        hashlib.md5(),
+    ]
+    for h in hashes:
+        h.update(blob)
+    return dict(
+        size_bytes=len(blob),
+        sha1hex=hashes[0].hexdigest(),
+        sha256hex=hashes[1].hexdigest(),
+        md5hex=hashes[2].hexdigest(),
+        mimetype=mimetype,
+    )
+
+def requests_retry_session(retries=2, backoff_factor=3,
+        status_forcelist=(500, 502, 504), session=None):
+    """
+    From: https://www.peterbe.com/plog/best-practice-with-retries-with-requests
+    """
+    session = session or requests.Session()
+    retry = Retry(
+        total=retries,
+        read=retries,
+        connect=retries,
+        backoff_factor=backoff_factor,
+        status_forcelist=status_forcelist,
+    )
+    adapter = HTTPAdapter(max_retries=retry)
+    session.mount('http://', adapter)
+    session.mount('https://', adapter)
+    return session
+
+def blob_path(sha1hex, directory="", file_suffix="", base_dir="."):
+    """
+    directory: eg, "png/"
+    sha1hex
+    file_suffix: eg, ".png"
+    """
+    fpath = "{}/{}{}/{}{}".format(
+            base_dir,
+            directory,
+            sha1hex[0:2],
+            sha1hex,
+            file_suffix)
+    return fpath
+
+def find_local_file(files, base_dir="."):
+    """
+    Takes a list of fatcat file entities (as dicts), and looks for a local file (PDF).
+
+    If none found, returns None.
+    If found, returns the file entity; the path can be determined from the sha1hex field.
+    """
+    for f in files:
+        if f['mimetype'] and not 'pdf' in f['mimetype'].lower():
+            continue
+        pdf_path = blob_path(f['sha1'], directory="pdf/", file_suffix=".pdf", base_dir=base_dir)
+        if os.path.isfile(pdf_path):
+            return copy.deepcopy(f)
+    return None
-- 
cgit v1.2.3