l: init nichtparasoup

2 files changed, 38 insertions, 0 deletions

diff --git a/lass/5pkgs/nichtparasoup/default.nix b/lass/5pkgs/nichtparasoup/default.nix
new file mode 100644
index 000000000..cf34c683f
--- /dev/null
+++ b/lass/5pkgs/nichtparasoup/default.nix
@@ -0,0 +1,25 @@
+{ stdenv, pkgs, ... }:
+let
+  py = pkgs.python3Packages.python.withPackages (p: [
+    p.werkzeug
+    p.beautifulsoup4
+  ]);
+  src = pkgs.fetchFromGitHub {
+    owner = "k4cg";
+    repo = "nichtparasoup";
+    rev = "cf164b5";
+    sha256 = "09bwh76agp14j8rv7bp47jcwhffc1b0bak0ikvzxyphph5lyidk9";
+  };
+  patchedSrc = stdenv.mkDerivation {
+    name = "nichtparasoup";
+    inherit src;
+    patches = [ ./exception.patch ];
+    phases = [ "unpackPhase" "patchPhase" "installPhase" ];
+    installPhase = ''
+      mkdir -p $out
+      cp -r * $out/
+    '';
+  };
+in pkgs.writeDashBin "nichtparasoup" ''
+  ${py}/bin/python ${patchedSrc}/nichtparasoup.py "$@"
+''
diff --git a/lass/5pkgs/nichtparasoup/exception.patch b/lass/5pkgs/nichtparasoup/exception.patch
new file mode 100644
index 000000000..34c177de0
--- /dev/null
+++ b/lass/5pkgs/nichtparasoup/exception.patch
@@ -0,0 +1,13 @@
+diff --git a/nichtparasoup.py b/nichtparasoup.py
+index 9da9a2b..833ca71 100755
+--- a/nichtparasoup.py
++++ b/nichtparasoup.py
+@@ -211,7 +211,7 @@ def cache_fill_loop():
+                     try:
+                         sources[crawler][site].crawl()
+                         info = Crawler.info()
+-                    except Exception, e:
++                    except Exception as e:
+                         logger.error("Error in crawler %s - %s: %s" % (crawler, site, e))
+                         break
+