backblogger/backblogger.py

# Backblogger.py : Scan directories for images, and scaffold into a blog post.
import os
import json
from datetime import datetime, date
import cv2
from PIL import Image as PILImage
from PIL.ExifTags import TAGS

def markdown_date(ts):
    d = datetime.fromtimestamp(ts)
    return f"{d.year}-{d.month}-{d.day}"

IMG_WIDTH = 760
class BlogImage:
    def __init__(self, filepath, article_number=None, image_number=None):
        self.path = filepath
        self.article_number = article_number
        self.image_number = image_number
        self._metadata = None
        self.resized = False
        self.file_date = os.stat(self.path).st_ctime if self.path else None

    def metadata(self, force_reload=False):
        if self._metadata and not force_reload: return self._metadata
        img = PILImage.open(self.path)
        exif = img.getexif()
        self._metadata = {TAGS.get(t, t): exif[t] for t in exif}
        return self._metadata

    def blog_image_name(self):
        AN = str(self.article_number).zfill(2)
        IN = str(self.image_number).zfill(2)
        EXT = os.path.basename(self.path).split('.')[1]
        return f"article{AN}_image{IN}.{EXT}"

    def resize(self, target_width, dest_fn=None):
        if dest_fn is None: dest_fn = self.blog_image_name()
        img = cv2.imread(self.path)
        h, w = img.shape[:2]
        if w <= target_width:
            cv2.imwrite(dest_fn, img)
            return dest_fn
        ratio = target_width / float(w)
        new_h = int(h * ratio)
        new_img = cv2.resize(img, (target_width, new_h), interpolation=cv2.INTER_AREA)
        cv2.imwrite(dest_fn, new_img)
        self.resized = True

    def markdown_template(self, template):
        img_replace = {
            "{{IMG_FN}}": self.blog_image_name(),
            "{{IMG_ORIG_FN}}": os.path.basename(self.path),
            "{{IMG_FILE_DATE}}": markdown_date(self.file_date),
            "{{IMG_META_DATE}}": self.metadata().get('DateTime', 'No Metadata')
            }
        for k in img_replace:
            template = template.replace(k, img_replace[k])
        return template

    def serialize(self):
        return {
            "path": os.path.basename(self.path),
            "metadata": self._metadata,
            "resized": self.resized,
            "resize_path": self.blog_image_name() if self.resized else '',
            "date": self.file_date
            }

    def deserialize(self, data, directory):
        self.path = os.path.join(directory, data['path'])
        self._metadata = data['metadata']
        self.resized = data['resized']
        if data['date']: os.stat(self.path).st_ctime if self.path else None

    def __repr__(self):
        return f"<BlogImage path={os.path.basename(self.path)}, article_number={self.article_number}, image_number={self.image_number}>"

    def best_date(self):
        meta_date = self.metadata().get('DateTime')
        if meta_date:
            try:
                return datetime.strptime(meta_date, '%Y:%m:%d %H:%M:%S').timestamp()
            except:
                print(f'DateTime exif failed parsing: {meta_date} ({self.path})')
        if self.file_date:
            return self.file_date
        return -1

    def __lt__(self, other):
        return self.best_date() < other.best_date()

class BlogScaffold:
    def __init__(self, path):
        self.path = os.path.abspath(path)
        self.data = { "images": [],
                      "blogfile": None
                    }
        self.blog_images = []
        # Check the path for backblog metadata
        self.scanned = os.path.exists(os.path.join(self.path, "backblog.json"))
        if not self.scanned:
            self.scan()
            return
        with open(os.path.join(self.path, "backblog.json"), "r") as f:
            self.data  = json.loads(f.read())

    def scan(self):
        _, _, files = next(os.walk(self.path))
        self.blog_images = []
        for f in files:
            if f.split('.')[1].lower() not in ('jpg', 'png', 'bmp', 'jpeg'): continue 
            self.blog_images.append(BlogImage(os.path.join(self.path, f)))
        self.scanned = True
        self.data['images'] = [bi.serialize() for bi in self.blog_images]
        self.save()
            
    def save(self):
        with open(os.path.join(self.path, "backblog.json"), "w") as f:
            f.write(json.dumps(self.data))

    def image_times(self):
        for i in self.data['images']:
            yield datetime.fromtimestamp(i['date'])

    def markdown_template(self, article_number):
        if not self.scanned: self.scan()
        replace = {
            "{{TITLE}}": "Backblog basic template about " + self.path,
            "{{SLUG}}": os.path.basename(self.path),
            "{{CATEGORY}}": "category",
            "{{EARLIESTDATE}}": markdown_date(min([i['date'] for i in self.data['images']])),
            "{{TODAY}}": str(date.today()),
            "{{ARTICLENUM}}": str(article_number).zfill(2)
            }
        txt = None
        with open("template.md", "r") as f:
            txt = f.read()
        img_template = txt.split("%%%")[1]
        img_txt = ""
        for i, image in enumerate(sorted(self.blog_images)):
            image.article_number = article_number
            image.image_number = i
            image.resize(IMG_WIDTH)
            img_txt += image.markdown_template(img_template)
        txt = txt.split("%%%")[0] + img_txt + txt.split("%%%")[2]
        for k in replace:
            txt = txt.replace(k, replace[k])
        template_fn = f"{str(article_number).zfill(2)}_{os.path.basename(self.path)}.md"
        render_template_fn = os.path.join(self.path, template_fn)
        with open(render_template_fn, "w") as f:
            f.write(txt)
        self.data['blogfile'] = template_fn
        

    def __repr__(self):
        if not self.scanned:
            return f"<BlogScaffold path={self.path}, scanned=False>"
        return f"<BlogScaffold path={self.path}, blogfile={self.data['blogfile']}, {len(self.data['images'])} image files>"

if __name__ == '__main__':
    subdirs = os.listdir('..')
    # don't scan program's directory
    subdirs.remove(os.path.basename(os.path.abspath('.')))
    #os.chdir('..')
    scaffolds = [BlogScaffold(os.path.join('..', sd)) for sd in subdirs]
    for s in scaffolds:
        s.scan()

    scaffolds.sort(key = lambda s: s.blog_images[-1], reverse=True)
    for s in scaffolds:
        print(f"{os.path.basename(s.path)} - earliest {markdown_date(s.blog_images[0].best_date())} - latest {markdown_date(s.blog_images[-1].best_date())}")