Added image convertion and page generation script

1 files changed, 245 insertions, 0 deletions

diff --git a/build.py b/build.py
new file mode 100755
index 0000000..f018b94
--- /dev/null
+++ b/build.py
@@ -0,0 +1,245 @@
+#!/usr/bin/env python3
+import jinja2
+import PIL.Image
+import PIL.ImageOps
+import argparse
+import yaml
+import os
+import re
+import hashlib
+import glob
+from datetime import datetime
+
+imagereg = re.compile("([a-z0-9]*)\\..*")
+
+
+def hashfile(fname):
+    md5 = hashlib.md5()
+    print(fname)
+    with open(fname, "rb") as f:
+        for chunk in iter(lambda: f.read(4096), b""):
+            md5.update(chunk)
+
+    return md5
+
+
+def parse_res(res):
+    return int(res)
+
+
+class FileLoader(jinja2.BaseLoader):
+    def get_source(self, environment, template):
+        if not os.path.exists(template):
+            raise jinja2.TemplateNotFound(template)
+        mtime = os.path.getmtime(template)
+        with open(template, "r") as f:
+            source = f.read()
+        return source, template, lambda: mtime == os.path.getmtime(template)
+
+
+class ImageLocation:
+    def __init__(self, folder):
+        self.folder = folder
+        self.stuff = {}
+
+        self.unused = set()
+        self.existing = {}
+
+        if not os.path.isdir(folder):
+            os.mkdir(folder)
+
+        self.look()
+
+        self.tmpname = os.path.join(folder, "tmp")
+
+    def look(self):
+        for name in os.listdir(self.folder):
+            m = imagereg.match(name)
+            if m is not None:
+                fhash = m.group(1)
+
+                self.unused.add(fhash)
+                self.existing[fhash] = name
+
+    def convert(self, imgname, settings):
+        # First determine the hash from settings
+        thash = hashlib.md5(
+                bytes(
+                    f"{settings['res']}+{settings['ext']}" + imgname,
+                    encoding="utf8")
+                ).hexdigest()
+
+        # Now check if it exists
+        if thash in self.existing:
+            print(f"Skipping file {imgname}[{thash}]")
+            return self.existing[thash], thash
+
+        # Okay convert it
+        fname = f"{thash}.{settings['ext']}"
+        print(f"Converting file {imgname} to {fname}")
+
+        tmpname = f"{self.tmpname}.{settings['ext']}"
+        with PIL.Image.open(imgname) as im:
+            target_height = settings["res"]
+            im = PIL.ImageOps.exif_transpose(im)
+            if im.size[0] > target_height:
+                res = (int((im.size[0] / im.size[1]) * target_height), target_height)
+                im = im.resize(res, PIL.Image.ANTIALIAS)
+
+            im.save(tmpname)
+
+        os.rename(tmpname, os.path.join(self.folder, fname))
+
+        self.existing[thash] = fname
+        self.unused.add(thash)
+
+        return fname, thash
+
+    def fetch(self, imgname, settings, relto=None):
+        # Check if we already have it
+        fname, hash = self.convert(imgname, settings)
+
+        # Mark the hash as used
+        self.unused.remove(hash)
+
+        if relto is not None:
+            full = os.path.join(self.folder, fname)
+            return os.path.relpath(full, start=relto)
+
+        return fname
+
+    def clean(self):
+        print("Running cleanup")
+        for uhash in self.unused:
+            fname = os.path.join(self.folder, self.existing[uhash])
+            print(f"Removing file {fname}")
+            os.remove(fname)
+
+class Image:
+    def __init__(self, filename):
+        self.name = os.path.basename(filename)
+        self.filename = filename
+        self.description = ""
+        self.scaledpath = ""
+        self.fullurl = ""
+
+        self.taken = None
+        self.load_metadata()
+
+    def load_metadata(self):
+        try:
+            with PIL.Image.open(self.filename) as im:
+                taken_str = im.getexif()[36867]
+        except KeyError:
+            print(f"Could not load date from image {self.filename}")
+            return
+
+        self.taken = datetime.strptime(taken_str, "%Y:%m:%d %H:%M:%S")
+
+
+def value_or_default(val, default=None):
+    if val is None:
+        return default
+    return val
+
+
+class Loader:
+    def __init__(self, config_file, loadpath):
+        self.config = Loader.__load_config(config_file)
+
+        self.loadpath = loadpath
+
+        self.images = []
+
+    def __load_config(config_file):
+        config = {}
+        with open(config_file, "r") as f:
+            config = yaml.safe_load(f)
+
+        print(f"Loaded config: {config}")
+        return config
+
+    def load(self, imagehook=None, clear=False):
+        # Clear if neccesary
+        if clear:
+            self.images = []
+
+        # First find all images from glob
+        for image in glob.iglob(os.path.join(self.loadpath, self.config["imageglob"])):
+            img = Image(image)
+            img.description = self.config["info"].get(img.name, "")
+
+            if imagehook is not None:
+                img = imagehook(img)
+
+            self.images.append(img)
+
+        self.images = sorted(self.images,
+                             key=lambda img: value_or_default(img.taken,
+                                                              datetime.min),
+                             reverse=True)
+
+
+class Renderer:
+    def __init__(self, loader, resolution, extension):
+        self.loader = loader
+        self.scaled = None
+
+        self.settings = {
+                "res": parse_res(resolution),
+                "ext": extension,
+                }
+
+    def build_to(self, dest, template, context, loc=None, clean=False):
+        # Load images
+        if loc is None:
+            loc = ImageLocation(os.path.join(dest, "imgs"))
+
+        def imgproc(img):
+            if context["cgit"] is not None:
+                img.fullurl = os.path.join(context["cgit"], "plain", img.name)
+
+            img.scaledpath = loc.fetch(img.filename, self.settings, relto=dest)
+
+            return img
+
+        self.loader.load(imgproc, clear=True)
+
+        jenv = jinja2.Environment(
+                loader=FileLoader(),
+                autoescape=jinja2.select_autoescape(['html', 'xml'])
+                )
+
+        tmpl = jenv.get_template(template)
+
+        with open(os.path.join(dest, "index.html"), "w") as f:
+            f.write(tmpl.render({
+                "ctx": context,
+                "images": self.loader.images
+                }))
+
+        if clean:
+            loc.clean()
+
+
+parser = argparse.ArgumentParser()
+parser.add_argument("--dest", "-d", default="build", help="where to put resulting files")
+parser.add_argument("--size", "-s", default="1080", help="size to scale web images to")
+parser.add_argument("--commit", "-g", help="git commit hash to announce")
+parser.add_argument("--clean",  help="clean unused image files", action="store_true")
+parser.add_argument("--config", "-c", default="imginfo.yml", help="where to load image definitions from")
+parser.add_argument("--template", "-t", default="index.html.j2", help="html template to use")
+parser.add_argument("--cgit", "-w", help="cgit repo base url")
+parser.add_argument("--load", "-l", default=".", help="where to load full size images from")
+parser.add_argument("--ext", "-e", default="png", help="image extension to use")
+
+args = parser.parse_args()
+
+context = {
+        "cgit": args.cgit,
+        "git": args.commit
+        }
+
+loader = Loader(args.config, args.load)
+rend = Renderer(loader, args.size, args.ext)
+rend.build_to(args.dest, args.template, context, clean=args.clean)