diff options
| -rw-r--r-- | Dockerfile | 40 | ||||
| -rw-r--r-- | README.md | 46 | ||||
| -rw-r--r-- | requirements.txt | 6 | ||||
| -rw-r--r-- | src/__legacy_src/app.nim (renamed from src/app.nim) | 2 | ||||
| -rw-r--r-- | src/__legacy_src/backend_compat.nim | 525 | ||||
| -rw-r--r-- | src/__legacy_src/server.nim | 525 | ||||
| -rw-r--r-- | src/app.py | 608 | ||||
| -rw-r--r-- | src/latex/template.tex | 88 | ||||
| -rw-r--r-- | src/templates/index.html | 2 |
9 files changed, 1729 insertions, 113 deletions
@@ -1,23 +1,37 @@ -FROM ubuntu:24.04 - -ENV DEBIAN_FRONTEND=noninteractive - -RUN apt-get update && apt-get install -y \ - nim \ - build-essential \ - pandoc \ - texlive-full \ +FROM python:3.12-slim + +ENV DEBIAN_FRONTEND=noninteractive \ + PYTHONDONTWRITEBYTECODE=1 \ + PYTHONUNBUFFERED=1 \ + PORT=5001 \ + HOST=0.0.0.0 \ + LOG_LEVEL=INFO \ + TRUST_PROXY=1 + +RUN apt-get update && apt-get install -y --no-install-recommends \ + libcairo2 \ + libpango-1.0-0 \ + libpangocairo-1.0-0 \ + libgdk-pixbuf-2.0-0 \ + shared-mime-info \ + fonts-noto \ fonts-noto-color-emoji \ && rm -rf /var/lib/apt/lists/* WORKDIR /app -COPY src/ . +RUN addgroup --system app && adduser --system --ingroup app app -RUN nim c -d:release --opt:size -o:likha-pdf app.nim +COPY requirements.txt . +RUN pip install --no-cache-dir --disable-pip-version-check -r requirements.txt + +COPY src/ . RUN mkdir -p generated uploads -EXPOSE 5000 +RUN chown -R app:app /app +USER app + +EXPOSE 5001 -CMD ["./likha-pdf"] +CMD ["gunicorn", "--bind", "0.0.0.0:5001", "--workers", "2", "--threads", "4", "--timeout", "180", "--graceful-timeout", "30", "--access-logfile", "-", "--error-logfile", "-", "app:app"] @@ -1,20 +1,23 @@ # likha-pdf -a simple and crappy web app that converts markdown to pdf using pandoc and lualatex. +a simple web app that converts markdown to pdf. + +<img width="1280" height="800" alt="likha-pdf screenshot" src="https://github.com/user-attachments/assets/578b8410-53ad-4323-a45a-19ae3aabe61d" /> -<img width="1440" height="1080" alt="likha-pdf screenshot" src="https://github.com/user-attachments/assets/9fa15803-f0da-43d0-8a82-0fc490f3d5ff" /> ## features - markdown to pdf export -- crappy image upload (but it works) -- emoji-capable latex template +- image upload with markdown snippet insertion +- paper size, margin, font, line spacing, and page number options +- syntax-highlighted code blocks +- always produces a pdf (reportlab fallback if weasyprint fails) ## requirements -- nim 1.6+ -- pandoc -- lualatex +- python 3.10+ +- system packages: `libcairo2 libpango-1.0-0 libpangocairo-1.0-0 libgdk-pixbuf2.0-0 shared-mime-info` +- gunicorn (installed from `requirements.txt`) ## image usage @@ -28,11 +31,34 @@ a simple and crappy web app that converts markdown to pdf using pandoc and luala ### local ```bash +python -m venv .venv +source .venv/bin/activate +pip install -r requirements.txt +cd src/ +python3 app.py +``` + +### production (vps + nginx) + +```bash cd src/ -nim c -d:release -o:likha-pdf app.nim -./likha-pdf +../.venv/bin/gunicorn \ + --bind 127.0.0.1:5001 \ + --workers 2 \ + --threads 4 \ + --timeout 180 \ + --graceful-timeout 30 \ + --access-logfile - \ + --error-logfile - \ + app:app ``` +nginx should reverse proxy to `127.0.0.1:5001` and pass: + +- `X-Forwarded-For` +- `X-Forwarded-Proto` +- `X-Forwarded-Host` + ### docker ```bash @@ -40,4 +66,4 @@ docker build -t likha-pdf . docker run -p 5001:5001 likha-pdf ``` -open `http://localhost:5000` +open `http://localhost:5001` diff --git a/requirements.txt b/requirements.txt new file mode 100644 index 0000000..e50d931 --- /dev/null +++ b/requirements.txt @@ -0,0 +1,6 @@ +flask==3.1.* +markdown==3.7.* +pygments==2.19.* +weasyprint==63.* +reportlab==4.3.* +gunicorn==23.0.*
\ No newline at end of file diff --git a/src/app.nim b/src/__legacy_src/app.nim index 17f95b7..8b64793 100644 --- a/src/app.nim +++ b/src/__legacy_src/app.nim @@ -522,4 +522,4 @@ when isMainModule: let server = newAsyncHttpServer() echo "listening on http://localhost:5001" - waitFor server.serve(Port(5001), route) + waitFor server.serve(Port(5001), route)
\ No newline at end of file diff --git a/src/__legacy_src/backend_compat.nim b/src/__legacy_src/backend_compat.nim new file mode 100644 index 0000000..8b64793 --- /dev/null +++ b/src/__legacy_src/backend_compat.nim @@ -0,0 +1,525 @@ +import + std/[ + asynchttpserver, asyncdispatch, os, osproc, streams, strutils, tables, times, uri, + random, + ] + +# tiny backend in nimlang, may be stupid, but this was fun + +const + AllowedImageExtensions = ["png", "jpg", "jpeg", "gif", "webp", "svg"] + ValidPaperSizes = [ + "a0paper", "a1paper", "a2paper", "a3paper", "a4paper", "a5paper", "a6paper", + "b0paper", "b1paper", "b2paper", "b3paper", "b4paper", "b5paper", "b6paper", + "c4paper", "c5paper", "c6paper", "letterpaper", "legalpaper", "executivepaper", + "ledgerpaper", "tabloid", "statement", "flsa", + ] + ValidMargins = ["0.25in", "0.5in", "0.75in", "1in", "1.25in", "1.5in", "1.75in"] + ValidLineSpacings = ["1", "1.5", "2"] + CustomPaperDimensions = [ + ("tabloid", "11in", "17in"), + ("statement", "5.5in", "8.5in"), + ("flsa", "8.5in", "13in"), + ] + +const AppName = "likha-pdf" + +proc lookupCustomPaper(name: string): tuple[width: string, height: string] = + for (paperName, w, h) in CustomPaperDimensions: + if paperName == name: + return (width: w, height: h) + (width: "", height: "") + +proc baseDir(): string {.inline.} = + getAppDir() + +proc generatedDir(): string {.inline.} = + baseDir() / "generated" + +proc uploadsDir(): string {.inline.} = + baseDir() / "uploads" + +proc latexTemplatePath(): string {.inline.} = + baseDir() / "latex" / "template.tex" + +proc templatesDir(): string {.inline.} = + baseDir() / "templates" + +proc partialsDir(): string {.inline.} = + templatesDir() / "partials" + +proc staticDir(): string {.inline.} = + baseDir() / "static" + +type MultipartPart = object + name: string + filename: string + contentType: string + content: string + +# helpers +proc htmlEscape(value: string): string = + result = value + result = result.replace("&", "&") + result = result.replace("<", "<") + result = result.replace(">", ">") + result = result.replace("\"", """) + result = result.replace("'", "'") + +proc randomHex(length: int): string = + const hexChars = "0123456789abcdef" + result = newStringOfCap(length) + for _ in 0 ..< length: + result.add(hexChars[rand(15)]) + +proc renderTemplate( + filePath: string, replacements: openArray[(string, string)] +): string = + result = readFile(filePath) + for (token, replacement) in replacements: + result = result.replace(token, replacement) + +proc decodeFormComponent(value: string): string = + decodeUrl(value.replace("+", " ")) + +proc parseUrlEncoded(body: string): Table[string, string] = + result = initTable[string, string]() + if body.len == 0: + return + + for pair in body.split("&"): + if pair.len == 0: + continue + let separator = pair.find('=') + if separator < 0: + result[decodeFormComponent(pair)] = "" + else: + let key = decodeFormComponent(pair[0 ..< separator]) + let value = decodeFormComponent(pair[separator + 1 .. ^1]) + result[key] = value + +# "options" are optional, defaults are forever. +proc pickOption(value: string, fallback: string, options: openArray[string]): string = + for option in options: + if option == value: + return value + fallback + +proc sanitizeFilename(filename: string): string = + result = newStringOfCap(filename.len) + for ch in filename: + if (ch >= 'a' and ch <= 'z') or (ch >= 'A' and ch <= 'Z') or + (ch >= '0' and ch <= '9') or (ch in {'-', '_', '.'}): + result.add(ch) + elif ch == ' ': + result.add('_') + +proc baseFilename(value: string): string = + var normalized = value.replace("\\", "/") + let index = normalized.rfind('/') + if index >= 0 and index < normalized.high: + normalized = normalized[index + 1 .. ^1] + elif index == normalized.high: + normalized = "" + normalized + +proc isAllowedImage(filename: string): bool = + let dot = filename.rfind('.') + if dot < 1 or dot == filename.high: + return false + let extension = filename[dot + 1 .. ^1].toLowerAscii() + for allowed in AllowedImageExtensions: + if extension == allowed: + return true + false + +proc tailText(value: string, maxLen: int = 1200): string = + if value.len <= maxLen: + return value + value[value.len - maxLen .. ^1] + +proc extractBoundary(contentType: string): string = + for part in contentType.split(';'): + let token = part.strip() + if token.toLowerAscii().startsWith("boundary="): + return token[9 .. ^1].strip(chars = {'\"', '\''}) + "" + +proc stripTrailingCrlf(value: string): string = + result = value + if result.len >= 2 and result.endsWith("\r\n"): + result.setLen(result.len - 2) + +# hand-rolled multipart parsing, yes i am aware that this is "eh" +proc parseMultipart(body: string, boundary: string): seq[MultipartPart] = + let delimiter = "--" & boundary + for rawChunk in body.split(delimiter): + var chunk = rawChunk + if chunk.len == 0: + continue + if chunk == "--" or chunk == "--\r\n": + continue + if chunk.startsWith("\r\n"): + chunk = chunk[2 .. ^1] + + chunk = stripTrailingCrlf(chunk) + + if chunk.len == 2 and chunk == "--": + continue + + let splitIndex = chunk.find("\r\n\r\n") + if splitIndex < 0: + continue + + let headerBlock = chunk[0 ..< splitIndex] + var content = chunk[splitIndex + 4 .. ^1] + content = stripTrailingCrlf(content) + + var name = "" + var filename = "" + var contentType = "application/octet-stream" + + for line in headerBlock.split("\r\n"): + let separator = line.find(':') + if separator <= 0: + continue + let headerName = line[0 ..< separator].strip().toLowerAscii() + let headerValue = line[separator + 1 .. ^1].strip() + + if headerName == "content-disposition": + for part in headerValue.split(';'): + let token = part.strip() + if token.startsWith("name="): + name = token[5 .. ^1].strip(chars = {'\"', '\''}) + elif token.startsWith("filename="): + filename = token[9 .. ^1].strip(chars = {'\"', '\''}) + elif headerName == "content-type": + contentType = headerValue + + if name.len > 0: + result.add( + MultipartPart( + name: name, filename: filename, contentType: contentType, content: content + ) + ) + +proc isSafeRelativePath(pathPart: string): bool = + pathPart.len > 0 and not pathPart.contains("..") and not pathPart.contains('\\') and + not pathPart.startsWith("/") + +proc fileContentType(filePath: string): string = + let lowered = filePath.toLowerAscii() + if lowered.endsWith(".js"): + return "application/javascript; charset=utf-8" + if lowered.endsWith(".css"): + return "text/css; charset=utf-8" + if lowered.endsWith(".html"): + return "text/html; charset=utf-8" + if lowered.endsWith(".png"): + return "image/png" + if lowered.endsWith(".jpg") or lowered.endsWith(".jpeg"): + return "image/jpeg" + if lowered.endsWith(".gif"): + return "image/gif" + if lowered.endsWith(".webp"): + return "image/webp" + if lowered.endsWith(".svg"): + return "image/svg+xml" + if lowered.endsWith(".pdf"): + return "application/pdf" + "application/octet-stream" + +# response wrappers +proc respondHtml(req: Request, code: HttpCode, content: string) {.async.} = + let headers = newHttpHeaders({"Content-Type": "text/html; charset=utf-8"}) + await req.respond(code, content, headers) + +proc respondText(req: Request, code: HttpCode, content: string) {.async.} = + let headers = newHttpHeaders({"Content-Type": "text/plain; charset=utf-8"}) + await req.respond(code, content, headers) + +proc respondFile( + req: Request, + filePath: string, + asAttachment: bool = false, + attachmentName: string = "", +) {.async.} = + if not fileExists(filePath): + await respondText(req, Http404, "Not found") + return + + var headers = newHttpHeaders() + headers["Content-Type"] = fileContentType(filePath) + if asAttachment and attachmentName.len > 0: + headers["Content-Disposition"] = "attachment; filename=\"" & attachmentName & "\"" + + await req.respond(Http200, readFile(filePath), headers) + +# pandoc does the heavy lifting +proc runPandoc( + sourceMarkdown: string, + outputPath: string, + paperSize: string, + margin: string, + mainFont: string, + lineSpacing: string, + showPageNumbers: bool, +): tuple[ok: bool, output: string, missingPandoc: bool] = + let tempDir = getTempDir() / (AppName & "-" & randomHex(10)) + createDir(tempDir) + let tempMarkdownPath = tempDir / "source.md" + let tempRawPath = tempDir / "raw.md" + + try: + # write raw markdown first + writeFile(tempRawPath, sourceMarkdown) + + # preprocess markdown: convert to ascii with transliteration and normalize quotes + let iconvCmd = + "iconv -c -t ASCII//TRANSLIT " & quoteShell(tempRawPath) & + " | sed 's/'\\''/'/g; s/\"\"/\"/g' > " & quoteShell(tempMarkdownPath) + let (_, iconvExitCode) = execCmdEx(iconvCmd) + + if iconvExitCode != 0: + # if preprocessing fails, fall back to original content + writeFile(tempMarkdownPath, sourceMarkdown) + + var args = @[ + tempMarkdownPath, + "--from", + "markdown+emoji+hard_line_breaks", + "--pdf-engine=lualatex", + "--template", + latexTemplatePath(), + "-V", + "margin=" & margin, + "-V", + "mainfont=" & mainFont, + "-V", + "linespacing=" & lineSpacing, + "--resource-path", + baseDir() & ":" & uploadsDir() & ":" & tempDir, + "-o", + outputPath, + ] + + let dims = lookupCustomPaper(paperSize) + if dims.width.len > 0: + args.add("-V") + args.add("paperwidth=" & dims.width) + args.add("-V") + args.add("paperheight=" & dims.height) + else: + args.add("-V") + args.add("papersize=" & paperSize) + + if not showPageNumbers: + args.add("-V") + args.add("hidepages=true") + + var process: Process + try: + process = + startProcess("pandoc", args = args, options = {poUsePath, poStdErrToStdOut}) + except OSError: + return ( + ok: false, + output: "Pandoc is not installed or not in PATH.", + missingPandoc: true, + ) + + let output = process.outputStream.readAll() + let exitCode = process.waitForExit() + process.close() + + if exitCode == 0: + return (ok: true, output: "", missingPandoc: false) + return (ok: false, output: output, missingPandoc: false) + finally: + try: + if fileExists(tempRawPath): + removeFile(tempRawPath) + if fileExists(tempMarkdownPath): + removeFile(tempMarkdownPath) + if dirExists(tempDir): + removeDir(tempDir) + except OSError: + discard + +# app endpoint: strict inputs, loud errors. +proc handleConvert(req: Request) {.async.} = + let formData = parseUrlEncoded(req.body) + let markdown = formData.getOrDefault("markdown", "").strip() + + if markdown.len == 0: + let html = renderTemplate( + partialsDir() / "error.html", [("{{ message }}", "Markdown content is required.")] + ) + await respondHtml(req, Http400, html) + return + + let paperSize = + pickOption(formData.getOrDefault("paper_size", ""), "a4paper", ValidPaperSizes) + let margin = pickOption(formData.getOrDefault("margin", ""), "1in", ValidMargins) + + var mainFontFamily = formData.getOrDefault("main_font", "serif") + if mainFontFamily != "serif" and mainFontFamily != "sans": + mainFontFamily = "serif" + + let mainFont = if mainFontFamily == "sans": "TeX Gyre Heros" else: "TeX Gyre Pagella" + let lineSpacing = + pickOption(formData.getOrDefault("line_spacing", ""), "1", ValidLineSpacings) + let showPageNumbers = formData.getOrDefault("page_numbers", "") == "on" + let epoch = int(getTime().toUnix()) + let outputName = AppName & "_" & $epoch & "_" & randomHex(32) & ".pdf" + let outputPath = generatedDir() / outputName + + let conversion = runPandoc( + markdown, outputPath, paperSize, margin, mainFont, lineSpacing, showPageNumbers + ) + + if not conversion.ok: + let message = + if conversion.missingPandoc: + conversion.output + else: + let stderr = conversion.output.strip() + if stderr.len > 0: + tailText(stderr) + else: + "PDF conversion failed." + + let html = renderTemplate( + partialsDir() / "error.html", [("{{ message }}", htmlEscape(message))] + ) + let code = if conversion.missingPandoc: Http500 else: Http400 + await respondHtml(req, code, html) + return + + let html = renderTemplate( + partialsDir() / "result.html", + [ + ("{{ filename }}", htmlEscape(outputName)), + ("{{ download_url }}", "/download/" & encodeUrl(outputName)), + ], + ) + await respondHtml(req, Http200, html) + +# upload endpoint. accepts image, returns markdown snippet +proc handleUploadImage(req: Request) {.async.} = + let contentType = req.headers.getOrDefault("Content-Type") + let boundary = extractBoundary(contentType) + + if boundary.len == 0: + let html = renderTemplate( + partialsDir() / "upload_error.html", + [("{{ message }}", "image file is required.")], + ) + await respondHtml(req, Http400, html) + return + + let parts = parseMultipart(req.body, boundary) + var imagePart: MultipartPart + var foundImage = false + for part in parts: + if part.name == "image": + imagePart = part + foundImage = true + break + + if not foundImage or imagePart.filename.strip().len == 0: + let html = renderTemplate( + partialsDir() / "upload_error.html", + [("{{ message }}", "image file is required.")], + ) + await respondHtml(req, Http400, html) + return + + let originalName = sanitizeFilename(baseFilename(imagePart.filename)) + if originalName.len == 0 or not isAllowedImage(originalName): + let html = renderTemplate( + partialsDir() / "upload_error.html", + [("{{ message }}", "unsupported image type.")], + ) + await respondHtml(req, Http400, html) + return + + let extensionStart = originalName.rfind('.') + let extension = originalName[extensionStart + 1 .. ^1].toLowerAscii() + + let epoch = int(getTime().toUnix()) + let storedName = "img_" & $epoch & "_" & randomHex(32) & "." & extension + let imagePath = uploadsDir() / storedName + + writeFile(imagePath, imagePart.content) + + let markdownSnippet = "![](uploads/" & storedName & ")" + let html = renderTemplate( + partialsDir() / "upload_result.html", + [ + ("{{ filename }}", htmlEscape(storedName)), + ("{{ markdown_snippet }}", htmlEscape(markdownSnippet)), + ("{{ preview_url }}", "/uploads/" & encodeUrl(storedName)), + ], + ) + await respondHtml(req, Http200, html) + +# router table +proc route(req: Request) {.async.} = + let path = req.url.path + + if req.reqMethod == HttpGet and path == "/": + await respondFile(req, templatesDir() / "index.html") + return + + if req.reqMethod == HttpGet and path.startsWith("/static/"): + let relativePath = decodeUrl(path[8 .. ^1]) + if not isSafeRelativePath(relativePath): + await respondText(req, Http400, "Invalid path") + return + await respondFile(req, staticDir() / relativePath) + return + + if req.reqMethod == HttpGet and path.startsWith("/uploads/"): + let relativePath = decodeUrl(path[9 .. ^1]) + if not isSafeRelativePath(relativePath): + await respondText(req, Http400, "Invalid path") + return + await respondFile(req, uploadsDir() / relativePath) + return + + if req.reqMethod == HttpGet and path.startsWith("/download/"): + let relativePath = decodeUrl(path[10 .. ^1]) + if not isSafeRelativePath(relativePath): + await respondText(req, Http400, "Invalid path") + return + await respondFile( + req, + generatedDir() / relativePath, + asAttachment = true, + attachmentName = relativePath, + ) + return + + if req.reqMethod == HttpPost and path == "/convert": + await handleConvert(req) + return + + if req.reqMethod == HttpPost and path == "/upload-image": + await handleUploadImage(req) + return + + await respondText(req, Http404, "Not found") + +# server boot, then we let htmx do htmx things. +when isMainModule: + randomize() + + if not dirExists(generatedDir()): + createDir(generatedDir()) + if not dirExists(uploadsDir()): + createDir(uploadsDir()) + + let server = newAsyncHttpServer() + echo "listening on http://localhost:5001" + waitFor server.serve(Port(5001), route)
\ No newline at end of file diff --git a/src/__legacy_src/server.nim b/src/__legacy_src/server.nim new file mode 100644 index 0000000..8b64793 --- /dev/null +++ b/src/__legacy_src/server.nim @@ -0,0 +1,525 @@ +import + std/[ + asynchttpserver, asyncdispatch, os, osproc, streams, strutils, tables, times, uri, + random, + ] + +# tiny backend in nimlang, may be stupid, but this was fun + +const + AllowedImageExtensions = ["png", "jpg", "jpeg", "gif", "webp", "svg"] + ValidPaperSizes = [ + "a0paper", "a1paper", "a2paper", "a3paper", "a4paper", "a5paper", "a6paper", + "b0paper", "b1paper", "b2paper", "b3paper", "b4paper", "b5paper", "b6paper", + "c4paper", "c5paper", "c6paper", "letterpaper", "legalpaper", "executivepaper", + "ledgerpaper", "tabloid", "statement", "flsa", + ] + ValidMargins = ["0.25in", "0.5in", "0.75in", "1in", "1.25in", "1.5in", "1.75in"] + ValidLineSpacings = ["1", "1.5", "2"] + CustomPaperDimensions = [ + ("tabloid", "11in", "17in"), + ("statement", "5.5in", "8.5in"), + ("flsa", "8.5in", "13in"), + ] + +const AppName = "likha-pdf" + +proc lookupCustomPaper(name: string): tuple[width: string, height: string] = + for (paperName, w, h) in CustomPaperDimensions: + if paperName == name: + return (width: w, height: h) + (width: "", height: "") + +proc baseDir(): string {.inline.} = + getAppDir() + +proc generatedDir(): string {.inline.} = + baseDir() / "generated" + +proc uploadsDir(): string {.inline.} = + baseDir() / "uploads" + +proc latexTemplatePath(): string {.inline.} = + baseDir() / "latex" / "template.tex" + +proc templatesDir(): string {.inline.} = + baseDir() / "templates" + +proc partialsDir(): string {.inline.} = + templatesDir() / "partials" + +proc staticDir(): string {.inline.} = + baseDir() / "static" + +type MultipartPart = object + name: string + filename: string + contentType: string + content: string + +# helpers +proc htmlEscape(value: string): string = + result = value + result = result.replace("&", "&") + result = result.replace("<", "<") + result = result.replace(">", ">") + result = result.replace("\"", """) + result = result.replace("'", "'") + +proc randomHex(length: int): string = + const hexChars = "0123456789abcdef" + result = newStringOfCap(length) + for _ in 0 ..< length: + result.add(hexChars[rand(15)]) + +proc renderTemplate( + filePath: string, replacements: openArray[(string, string)] +): string = + result = readFile(filePath) + for (token, replacement) in replacements: + result = result.replace(token, replacement) + +proc decodeFormComponent(value: string): string = + decodeUrl(value.replace("+", " ")) + +proc parseUrlEncoded(body: string): Table[string, string] = + result = initTable[string, string]() + if body.len == 0: + return + + for pair in body.split("&"): + if pair.len == 0: + continue + let separator = pair.find('=') + if separator < 0: + result[decodeFormComponent(pair)] = "" + else: + let key = decodeFormComponent(pair[0 ..< separator]) + let value = decodeFormComponent(pair[separator + 1 .. ^1]) + result[key] = value + +# "options" are optional, defaults are forever. +proc pickOption(value: string, fallback: string, options: openArray[string]): string = + for option in options: + if option == value: + return value + fallback + +proc sanitizeFilename(filename: string): string = + result = newStringOfCap(filename.len) + for ch in filename: + if (ch >= 'a' and ch <= 'z') or (ch >= 'A' and ch <= 'Z') or + (ch >= '0' and ch <= '9') or (ch in {'-', '_', '.'}): + result.add(ch) + elif ch == ' ': + result.add('_') + +proc baseFilename(value: string): string = + var normalized = value.replace("\\", "/") + let index = normalized.rfind('/') + if index >= 0 and index < normalized.high: + normalized = normalized[index + 1 .. ^1] + elif index == normalized.high: + normalized = "" + normalized + +proc isAllowedImage(filename: string): bool = + let dot = filename.rfind('.') + if dot < 1 or dot == filename.high: + return false + let extension = filename[dot + 1 .. ^1].toLowerAscii() + for allowed in AllowedImageExtensions: + if extension == allowed: + return true + false + +proc tailText(value: string, maxLen: int = 1200): string = + if value.len <= maxLen: + return value + value[value.len - maxLen .. ^1] + +proc extractBoundary(contentType: string): string = + for part in contentType.split(';'): + let token = part.strip() + if token.toLowerAscii().startsWith("boundary="): + return token[9 .. ^1].strip(chars = {'\"', '\''}) + "" + +proc stripTrailingCrlf(value: string): string = + result = value + if result.len >= 2 and result.endsWith("\r\n"): + result.setLen(result.len - 2) + +# hand-rolled multipart parsing, yes i am aware that this is "eh" +proc parseMultipart(body: string, boundary: string): seq[MultipartPart] = + let delimiter = "--" & boundary + for rawChunk in body.split(delimiter): + var chunk = rawChunk + if chunk.len == 0: + continue + if chunk == "--" or chunk == "--\r\n": + continue + if chunk.startsWith("\r\n"): + chunk = chunk[2 .. ^1] + + chunk = stripTrailingCrlf(chunk) + + if chunk.len == 2 and chunk == "--": + continue + + let splitIndex = chunk.find("\r\n\r\n") + if splitIndex < 0: + continue + + let headerBlock = chunk[0 ..< splitIndex] + var content = chunk[splitIndex + 4 .. ^1] + content = stripTrailingCrlf(content) + + var name = "" + var filename = "" + var contentType = "application/octet-stream" + + for line in headerBlock.split("\r\n"): + let separator = line.find(':') + if separator <= 0: + continue + let headerName = line[0 ..< separator].strip().toLowerAscii() + let headerValue = line[separator + 1 .. ^1].strip() + + if headerName == "content-disposition": + for part in headerValue.split(';'): + let token = part.strip() + if token.startsWith("name="): + name = token[5 .. ^1].strip(chars = {'\"', '\''}) + elif token.startsWith("filename="): + filename = token[9 .. ^1].strip(chars = {'\"', '\''}) + elif headerName == "content-type": + contentType = headerValue + + if name.len > 0: + result.add( + MultipartPart( + name: name, filename: filename, contentType: contentType, content: content + ) + ) + +proc isSafeRelativePath(pathPart: string): bool = + pathPart.len > 0 and not pathPart.contains("..") and not pathPart.contains('\\') and + not pathPart.startsWith("/") + +proc fileContentType(filePath: string): string = + let lowered = filePath.toLowerAscii() + if lowered.endsWith(".js"): + return "application/javascript; charset=utf-8" + if lowered.endsWith(".css"): + return "text/css; charset=utf-8" + if lowered.endsWith(".html"): + return "text/html; charset=utf-8" + if lowered.endsWith(".png"): + return "image/png" + if lowered.endsWith(".jpg") or lowered.endsWith(".jpeg"): + return "image/jpeg" + if lowered.endsWith(".gif"): + return "image/gif" + if lowered.endsWith(".webp"): + return "image/webp" + if lowered.endsWith(".svg"): + return "image/svg+xml" + if lowered.endsWith(".pdf"): + return "application/pdf" + "application/octet-stream" + +# response wrappers +proc respondHtml(req: Request, code: HttpCode, content: string) {.async.} = + let headers = newHttpHeaders({"Content-Type": "text/html; charset=utf-8"}) + await req.respond(code, content, headers) + +proc respondText(req: Request, code: HttpCode, content: string) {.async.} = + let headers = newHttpHeaders({"Content-Type": "text/plain; charset=utf-8"}) + await req.respond(code, content, headers) + +proc respondFile( + req: Request, + filePath: string, + asAttachment: bool = false, + attachmentName: string = "", +) {.async.} = + if not fileExists(filePath): + await respondText(req, Http404, "Not found") + return + + var headers = newHttpHeaders() + headers["Content-Type"] = fileContentType(filePath) + if asAttachment and attachmentName.len > 0: + headers["Content-Disposition"] = "attachment; filename=\"" & attachmentName & "\"" + + await req.respond(Http200, readFile(filePath), headers) + +# pandoc does the heavy lifting +proc runPandoc( + sourceMarkdown: string, + outputPath: string, + paperSize: string, + margin: string, + mainFont: string, + lineSpacing: string, + showPageNumbers: bool, +): tuple[ok: bool, output: string, missingPandoc: bool] = + let tempDir = getTempDir() / (AppName & "-" & randomHex(10)) + createDir(tempDir) + let tempMarkdownPath = tempDir / "source.md" + let tempRawPath = tempDir / "raw.md" + + try: + # write raw markdown first + writeFile(tempRawPath, sourceMarkdown) + + # preprocess markdown: convert to ascii with transliteration and normalize quotes + let iconvCmd = + "iconv -c -t ASCII//TRANSLIT " & quoteShell(tempRawPath) & + " | sed 's/'\\''/'/g; s/\"\"/\"/g' > " & quoteShell(tempMarkdownPath) + let (_, iconvExitCode) = execCmdEx(iconvCmd) + + if iconvExitCode != 0: + # if preprocessing fails, fall back to original content + writeFile(tempMarkdownPath, sourceMarkdown) + + var args = @[ + tempMarkdownPath, + "--from", + "markdown+emoji+hard_line_breaks", + "--pdf-engine=lualatex", + "--template", + latexTemplatePath(), + "-V", + "margin=" & margin, + "-V", + "mainfont=" & mainFont, + "-V", + "linespacing=" & lineSpacing, + "--resource-path", + baseDir() & ":" & uploadsDir() & ":" & tempDir, + "-o", + outputPath, + ] + + let dims = lookupCustomPaper(paperSize) + if dims.width.len > 0: + args.add("-V") + args.add("paperwidth=" & dims.width) + args.add("-V") + args.add("paperheight=" & dims.height) + else: + args.add("-V") + args.add("papersize=" & paperSize) + + if not showPageNumbers: + args.add("-V") + args.add("hidepages=true") + + var process: Process + try: + process = + startProcess("pandoc", args = args, options = {poUsePath, poStdErrToStdOut}) + except OSError: + return ( + ok: false, + output: "Pandoc is not installed or not in PATH.", + missingPandoc: true, + ) + + let output = process.outputStream.readAll() + let exitCode = process.waitForExit() + process.close() + + if exitCode == 0: + return (ok: true, output: "", missingPandoc: false) + return (ok: false, output: output, missingPandoc: false) + finally: + try: + if fileExists(tempRawPath): + removeFile(tempRawPath) + if fileExists(tempMarkdownPath): + removeFile(tempMarkdownPath) + if dirExists(tempDir): + removeDir(tempDir) + except OSError: + discard + +# app endpoint: strict inputs, loud errors. +proc handleConvert(req: Request) {.async.} = + let formData = parseUrlEncoded(req.body) + let markdown = formData.getOrDefault("markdown", "").strip() + + if markdown.len == 0: + let html = renderTemplate( + partialsDir() / "error.html", [("{{ message }}", "Markdown content is required.")] + ) + await respondHtml(req, Http400, html) + return + + let paperSize = + pickOption(formData.getOrDefault("paper_size", ""), "a4paper", ValidPaperSizes) + let margin = pickOption(formData.getOrDefault("margin", ""), "1in", ValidMargins) + + var mainFontFamily = formData.getOrDefault("main_font", "serif") + if mainFontFamily != "serif" and mainFontFamily != "sans": + mainFontFamily = "serif" + + let mainFont = if mainFontFamily == "sans": "TeX Gyre Heros" else: "TeX Gyre Pagella" + let lineSpacing = + pickOption(formData.getOrDefault("line_spacing", ""), "1", ValidLineSpacings) + let showPageNumbers = formData.getOrDefault("page_numbers", "") == "on" + let epoch = int(getTime().toUnix()) + let outputName = AppName & "_" & $epoch & "_" & randomHex(32) & ".pdf" + let outputPath = generatedDir() / outputName + + let conversion = runPandoc( + markdown, outputPath, paperSize, margin, mainFont, lineSpacing, showPageNumbers + ) + + if not conversion.ok: + let message = + if conversion.missingPandoc: + conversion.output + else: + let stderr = conversion.output.strip() + if stderr.len > 0: + tailText(stderr) + else: + "PDF conversion failed." + + let html = renderTemplate( + partialsDir() / "error.html", [("{{ message }}", htmlEscape(message))] + ) + let code = if conversion.missingPandoc: Http500 else: Http400 + await respondHtml(req, code, html) + return + + let html = renderTemplate( + partialsDir() / "result.html", + [ + ("{{ filename }}", htmlEscape(outputName)), + ("{{ download_url }}", "/download/" & encodeUrl(outputName)), + ], + ) + await respondHtml(req, Http200, html) + +# upload endpoint. accepts image, returns markdown snippet +proc handleUploadImage(req: Request) {.async.} = + let contentType = req.headers.getOrDefault("Content-Type") + let boundary = extractBoundary(contentType) + + if boundary.len == 0: + let html = renderTemplate( + partialsDir() / "upload_error.html", + [("{{ message }}", "image file is required.")], + ) + await respondHtml(req, Http400, html) + return + + let parts = parseMultipart(req.body, boundary) + var imagePart: MultipartPart + var foundImage = false + for part in parts: + if part.name == "image": + imagePart = part + foundImage = true + break + + if not foundImage or imagePart.filename.strip().len == 0: + let html = renderTemplate( + partialsDir() / "upload_error.html", + [("{{ message }}", "image file is required.")], + ) + await respondHtml(req, Http400, html) + return + + let originalName = sanitizeFilename(baseFilename(imagePart.filename)) + if originalName.len == 0 or not isAllowedImage(originalName): + let html = renderTemplate( + partialsDir() / "upload_error.html", + [("{{ message }}", "unsupported image type.")], + ) + await respondHtml(req, Http400, html) + return + + let extensionStart = originalName.rfind('.') + let extension = originalName[extensionStart + 1 .. ^1].toLowerAscii() + + let epoch = int(getTime().toUnix()) + let storedName = "img_" & $epoch & "_" & randomHex(32) & "." & extension + let imagePath = uploadsDir() / storedName + + writeFile(imagePath, imagePart.content) + + let markdownSnippet = "![](uploads/" & storedName & ")" + let html = renderTemplate( + partialsDir() / "upload_result.html", + [ + ("{{ filename }}", htmlEscape(storedName)), + ("{{ markdown_snippet }}", htmlEscape(markdownSnippet)), + ("{{ preview_url }}", "/uploads/" & encodeUrl(storedName)), + ], + ) + await respondHtml(req, Http200, html) + +# router table +proc route(req: Request) {.async.} = + let path = req.url.path + + if req.reqMethod == HttpGet and path == "/": + await respondFile(req, templatesDir() / "index.html") + return + + if req.reqMethod == HttpGet and path.startsWith("/static/"): + let relativePath = decodeUrl(path[8 .. ^1]) + if not isSafeRelativePath(relativePath): + await respondText(req, Http400, "Invalid path") + return + await respondFile(req, staticDir() / relativePath) + return + + if req.reqMethod == HttpGet and path.startsWith("/uploads/"): + let relativePath = decodeUrl(path[9 .. ^1]) + if not isSafeRelativePath(relativePath): + await respondText(req, Http400, "Invalid path") + return + await respondFile(req, uploadsDir() / relativePath) + return + + if req.reqMethod == HttpGet and path.startsWith("/download/"): + let relativePath = decodeUrl(path[10 .. ^1]) + if not isSafeRelativePath(relativePath): + await respondText(req, Http400, "Invalid path") + return + await respondFile( + req, + generatedDir() / relativePath, + asAttachment = true, + attachmentName = relativePath, + ) + return + + if req.reqMethod == HttpPost and path == "/convert": + await handleConvert(req) + return + + if req.reqMethod == HttpPost and path == "/upload-image": + await handleUploadImage(req) + return + + await respondText(req, Http404, "Not found") + +# server boot, then we let htmx do htmx things. +when isMainModule: + randomize() + + if not dirExists(generatedDir()): + createDir(generatedDir()) + if not dirExists(uploadsDir()): + createDir(uploadsDir()) + + let server = newAsyncHttpServer() + echo "listening on http://localhost:5001" + waitFor server.serve(Port(5001), route)
\ No newline at end of file diff --git a/src/app.py b/src/app.py new file mode 100644 index 0000000..89666ad --- /dev/null +++ b/src/app.py @@ -0,0 +1,608 @@ +#!/usr/bin/env python3 + +# likha-pdf — markdown to pdf, no latex required +# production-friendly flask app with weasyprint + reportlab fallback + +import logging +import os +import secrets +import time +from pathlib import Path, PurePosixPath + +from flask import ( + Flask, + Response, + current_app, + request, + send_from_directory, + abort, +) +from markupsafe import escape +from markdown import markdown +from weasyprint import HTML +from werkzeug.middleware.proxy_fix import ProxyFix + +APP_NAME = "likha-pdf" +DEFAULT_HOST = "0.0.0.0" +DEFAULT_PORT = 5001 + +BASE_DIR = Path(__file__).resolve().parent +GENERATED_DIR = BASE_DIR / "generated" +UPLOADS_DIR = BASE_DIR / "uploads" +TEMPLATES_DIR = BASE_DIR / "templates" +PARTIALS_DIR = TEMPLATES_DIR / "partials" +STATIC_DIR = BASE_DIR / "static" + +ALLOWED_IMAGE_EXTS = {"png", "jpg", "jpeg", "gif", "webp", "svg"} + +VALID_PAPER_SIZES = { + "a0paper", "a1paper", "a2paper", "a3paper", "a4paper", "a5paper", "a6paper", + "b0paper", "b1paper", "b2paper", "b3paper", "b4paper", "b5paper", "b6paper", + "c4paper", "c5paper", "c6paper", + "letterpaper", "legalpaper", "executivepaper", + "ledgerpaper", "tabloid", "statement", "flsa", +} + +VALID_MARGINS = {"0.25in", "0.5in", "0.75in", "1in", "1.25in", "1.5in", "1.75in"} + +VALID_LINE_SPACINGS = {"1", "1.5", "2"} + +# css page dimensions for each paper size +PAPER_CSS = { + "a0paper": "841mm 1189mm", + "a1paper": "594mm 841mm", + "a2paper": "420mm 594mm", + "a3paper": "297mm 420mm", + "a4paper": "210mm 297mm", + "a5paper": "148mm 210mm", + "a6paper": "105mm 148mm", + "b0paper": "1000mm 1414mm", + "b1paper": "707mm 1000mm", + "b2paper": "500mm 707mm", + "b3paper": "353mm 500mm", + "b4paper": "250mm 353mm", + "b5paper": "176mm 250mm", + "b6paper": "125mm 176mm", + "c4paper": "229mm 324mm", + "c5paper": "162mm 229mm", + "c6paper": "114mm 162mm", + "letterpaper": "8.5in 11in", + "legalpaper": "8.5in 14in", + "executivepaper": "7in 10in", + "ledgerpaper": "17in 11in", + "tabloid": "11in 17in", + "statement": "5.5in 8.5in", + "flsa": "8.5in 13in", +} + +MARKDOWN_EXTENSIONS = [ + "tables", + "fenced_code", + "codehilite", + "nl2br", + "sane_lists", + "smarty", + "toc", + "attr_list", + "md_in_html", +] + +MARKDOWN_EXT_CONFIG = { + "codehilite": { + "css_class": "highlight", + "guess_lang": True, + "noclasses": True, + }, +} + + +# helpers +def env_bool(name, default=False): + raw = os.getenv(name) + if raw is None: + return default + return raw.strip().lower() in {"1", "true", "yes", "on"} + + +def ensure_runtime_dirs(): + GENERATED_DIR.mkdir(parents=True, exist_ok=True) + UPLOADS_DIR.mkdir(parents=True, exist_ok=True) + + +def random_hex(length=32): + return secrets.token_hex(length // 2) + + +def pick_option(value, fallback, valid): + return value if value in valid else fallback + + +def sanitize_filename(name): + """keep only safe characters in a filename""" + name = os.path.basename(name.replace("\\", "/")) + out = [] + for ch in name: + if ch.isalnum() or ch in "-_.": + out.append(ch) + elif ch == " ": + out.append("_") + return "".join(out) + + +def is_allowed_image(filename): + dot = filename.rfind(".") + if dot < 1 or dot == len(filename) - 1: + return False + ext = filename[dot + 1:].lower() + return ext in ALLOWED_IMAGE_EXTS + + +def is_safe_relative_path(path_part): + if not path_part or "\\" in path_part: + return False + safe_path = PurePosixPath(path_part) + return not safe_path.is_absolute() and ".." not in safe_path.parts + + +def read_partial(name, replacements=None): + """read a partial html template and apply replacements""" + content = (PARTIALS_DIR / name).read_text(encoding="utf-8") + if replacements: + for token, value in replacements.items(): + content = content.replace(token, value) + return content + + +def tail_text(value, max_len=1200): + if len(value) <= max_len: + return value + return value[-max_len:] + + +# pdf stylesheet generator +def build_pdf_css(paper_size, margin, font_family, line_spacing, show_page_numbers): + """build the css for weasyprint pdf rendering""" + page_dims = PAPER_CSS.get(paper_size, "8.5in 11in") + + if font_family == "sans": + font_stack = '"Helvetica Neue", Helvetica, Arial, "Noto Sans", sans-serif' + else: + font_stack = '"Georgia", "Noto Serif", "Times New Roman", serif' + + page_number_css = "" + if show_page_numbers: + page_number_css = """ + @bottom-center { + content: counter(page); + font-size: 9pt; + color: #666; + }""" + + return f""" +@page {{ + size: {page_dims}; + margin: {margin};{page_number_css} +}} + +body {{ + font-family: {font_stack}; + font-size: 11pt; + line-height: {line_spacing}; + color: #000; + word-wrap: break-word; + overflow-wrap: break-word; +}} + +h1, h2, h3, h4, h5, h6 {{ + margin-top: 1em; + margin-bottom: 0.4em; + page-break-after: avoid; +}} + +h1 {{ font-size: 20pt; }} +h2 {{ font-size: 16pt; }} +h3 {{ font-size: 13pt; }} +h4 {{ font-size: 11pt; }} + +p {{ + margin: 0 0 0.6em 0; +}} + +pre {{ + background: #f5f5f5; + border: 1px solid #ddd; + border-radius: 3px; + padding: 0.6em; + font-size: 9pt; + white-space: pre-wrap; + word-wrap: break-word; + overflow-wrap: break-word; + page-break-inside: avoid; +}} + +code {{ + font-family: "Courier New", Courier, "Liberation Mono", monospace; + font-size: 9pt; +}} + +p > code, li > code {{ + background: #f0f0f0; + padding: 0.1em 0.3em; + border-radius: 2px; +}} + +blockquote {{ + border-left: 3px solid #ccc; + margin: 0.6em 0; + padding: 0.3em 0.8em; + color: #555; +}} + +table {{ + border-collapse: collapse; + width: 100%; + margin: 0.6em 0; + page-break-inside: avoid; +}} + +th, td {{ + border: 1px solid #ccc; + padding: 0.4em 0.6em; + text-align: left; +}} + +th {{ + background: #f5f5f5; + font-weight: bold; +}} + +img {{ + max-width: 100%; + height: auto; +}} + +a {{ + color: #0066cc; + text-decoration: underline; +}} + +hr {{ + border: none; + border-top: 1px solid #ccc; + margin: 1em 0; +}} + +ul, ol {{ + margin: 0.4em 0; + padding-left: 1.5em; +}} + +li {{ + margin-bottom: 0.2em; +}} +""" + + +# pdf conversion +def markdown_to_html(source): + """convert markdown text to an html fragment""" + return markdown( + source, + extensions=MARKDOWN_EXTENSIONS, + extension_configs=MARKDOWN_EXT_CONFIG, + ) + + +def build_full_html(body_html, css): + """wrap the converted html body in a full document with styles""" + return f"""<!doctype html> +<html> +<head> +<meta charset="utf-8" /> +<style> +{css} +</style> +</head> +<body> +{body_html} +</body> +</html>""" + + +def convert_with_weasyprint(full_html, output_path): + """render html to pdf via weasyprint. returns (ok, error_msg).""" + try: + doc = HTML( + string=full_html, + base_url=str(BASE_DIR), + ) + doc.write_pdf(output_path) + return True, "" + except Exception as exc: + return False, str(exc) + + +def convert_with_reportlab(source_markdown, output_path, paper_size, margin, + font_family, line_spacing): + """fallback: produce a basic text pdf with reportlab. + not pretty, but guarantees a file is always created.""" + from reportlab.lib.pagesizes import ( + A0, A1, A2, A3, A4, A5, A6, + B0, B1, B2, B3, B4, B5, B6, + LETTER, LEGAL, LEDGER, TABLOID, + ) + from reportlab.lib.units import inch, mm + from reportlab.platypus import SimpleDocTemplate, Paragraph, Spacer, Preformatted + from reportlab.lib.styles import getSampleStyleSheet, ParagraphStyle + from reportlab.lib.enums import TA_LEFT + + size_map = { + "a0paper": A0, "a1paper": A1, "a2paper": A2, "a3paper": A3, + "a4paper": A4, "a5paper": A5, "a6paper": A6, + "b0paper": B0, "b1paper": B1, "b2paper": B2, "b3paper": B3, + "b4paper": B4, "b5paper": B5, "b6paper": B6, + "letterpaper": LETTER, "legalpaper": LEGAL, + "executivepaper": (7 * inch, 10 * inch), + "ledgerpaper": LEDGER, "tabloid": TABLOID, + "statement": (5.5 * inch, 8.5 * inch), + "flsa": (8.5 * inch, 13 * inch), + "c4paper": (229 * mm, 324 * mm), + "c5paper": (162 * mm, 229 * mm), + "c6paper": (114 * mm, 162 * mm), + } + + margin_map = { + "0.25in": 0.25 * inch, "0.5in": 0.5 * inch, "0.75in": 0.75 * inch, + "1in": 1.0 * inch, "1.25in": 1.25 * inch, "1.5in": 1.5 * inch, + "1.75in": 1.75 * inch, + } + + pagesize = size_map.get(paper_size, LETTER) + m = margin_map.get(margin, 1.0 * inch) + + doc = SimpleDocTemplate( + output_path, + pagesize=pagesize, + leftMargin=m, rightMargin=m, + topMargin=m, bottomMargin=m, + ) + + styles = getSampleStyleSheet() + font_name = "Helvetica" if font_family == "sans" else "Times-Roman" + spacing_val = float(line_spacing) if line_spacing else 1.0 + + body_style = ParagraphStyle( + "BodyCustom", + parent=styles["Normal"], + fontName=font_name, + fontSize=11, + leading=11 * spacing_val * 1.2, + alignment=TA_LEFT, + ) + + code_style = ParagraphStyle( + "CodeCustom", + parent=styles["Code"], + fontName="Courier", + fontSize=9, + leading=11, + leftIndent=12, + ) + + story = [] + in_code_block = False + code_lines = [] + + for line in source_markdown.splitlines(): + if line.startswith("```"): + if in_code_block: + # close code block + code_text = "\n".join(code_lines) + story.append(Preformatted(code_text, code_style)) + story.append(Spacer(1, 6)) + code_lines = [] + in_code_block = False + else: + in_code_block = True + continue + + if in_code_block: + code_lines.append(line) + continue + + stripped = line.strip() + + if not stripped: + story.append(Spacer(1, 6)) + continue + + # heading detection + if stripped.startswith("#"): + level = len(stripped) - len(stripped.lstrip("#")) + level = min(level, 6) + text = stripped.lstrip("#").strip() + heading_style = ParagraphStyle( + f"H{level}", + parent=styles["Heading1"], + fontName=font_name, + fontSize=max(20 - (level * 2), 11), + ) + story.append(Paragraph(text, heading_style)) + story.append(Spacer(1, 4)) + continue + + story.append(Paragraph(line, body_style)) + + # flush any unclosed code block + if code_lines: + code_text = "\n".join(code_lines) + story.append(Preformatted(code_text, code_style)) + + doc.build(story) + + +def generate_pdf(source_markdown, output_path, paper_size, margin, + font_family, line_spacing, show_page_numbers): + """convert markdown to pdf. always produces a file.""" + body_html = markdown_to_html(source_markdown) + css = build_pdf_css(paper_size, margin, font_family, line_spacing, show_page_numbers) + full_html = build_full_html(body_html, css) + + ok, err = convert_with_weasyprint(full_html, output_path) + if ok: + return True, "" + + # weasyprint failed — fall back to reportlab + try: + current_app.logger.warning("weasyprint failed, using reportlab fallback: %s", err) + convert_with_reportlab( + source_markdown, output_path, + paper_size, margin, font_family, line_spacing, + ) + return True, f"(used fallback renderer) {err}" + except Exception as fallback_err: + return False, f"weasyprint: {err} | reportlab: {fallback_err}" + + +def create_app(): + ensure_runtime_dirs() + + app = Flask( + __name__, + template_folder=str(TEMPLATES_DIR), + static_folder=str(STATIC_DIR), + static_url_path="/static", + ) + + app.config["MAX_CONTENT_LENGTH"] = int(os.getenv("MAX_CONTENT_LENGTH", str(64 * 1024 * 1024))) + + if env_bool("TRUST_PROXY", default=True): + app.wsgi_app = ProxyFix(app.wsgi_app, x_for=1, x_proto=1, x_host=1, x_port=1) + + log_level = os.getenv("LOG_LEVEL", "INFO").upper() + app.logger.setLevel(log_level) + + @app.after_request + def add_security_headers(resp): + resp.headers.setdefault("X-Content-Type-Options", "nosniff") + resp.headers.setdefault("X-Frame-Options", "DENY") + resp.headers.setdefault("Referrer-Policy", "no-referrer") + return resp + + @app.errorhandler(413) + def payload_too_large(_err): + return read_partial("upload_error.html", { + "{{ message }}": "request body too large.", + }), 413 + + @app.route("/healthz") + def healthz(): + return Response("ok\n", mimetype="text/plain") + + @app.route("/") + def index(): + return send_from_directory(str(TEMPLATES_DIR), "index.html") + + @app.route("/convert", methods=["POST"]) + def convert(): + md = request.form.get("markdown", "").strip() + if not md: + return read_partial("error.html", { + "{{ message }}": "Markdown content is required.", + }), 400 + + paper_size = pick_option( + request.form.get("paper_size", ""), "letterpaper", VALID_PAPER_SIZES, + ) + margin = pick_option( + request.form.get("margin", ""), "1in", VALID_MARGINS, + ) + + font_family = request.form.get("main_font", "serif") + if font_family not in ("serif", "sans"): + font_family = "serif" + + line_spacing = pick_option( + request.form.get("line_spacing", ""), "1", VALID_LINE_SPACINGS, + ) + show_page_numbers = request.form.get("page_numbers") == "on" + + output_name = f"{APP_NAME}_{int(time.time())}_{random_hex()}.pdf" + output_path = GENERATED_DIR / output_name + + ok, err = generate_pdf( + md, + str(output_path), + paper_size, + margin, + font_family, + line_spacing, + show_page_numbers, + ) + + if not ok: + app.logger.error("pdf generation failed: %s", err) + return read_partial("error.html", { + "{{ message }}": str(escape(tail_text(err))), + }), 500 + + return read_partial("result.html", { + "{{ filename }}": str(escape(output_name)), + "{{ download_url }}": f"/download/{output_name}", + }) + + @app.route("/upload-image", methods=["POST"]) + def upload_image(): + uploaded = request.files.get("image") + if not uploaded or not uploaded.filename or not uploaded.filename.strip(): + return read_partial("upload_error.html", { + "{{ message }}": "image file is required.", + }), 400 + + original = sanitize_filename(uploaded.filename) + if not original or not is_allowed_image(original): + return read_partial("upload_error.html", { + "{{ message }}": "unsupported image type.", + }), 400 + + ext = original.rsplit(".", 1)[-1].lower() + stored_name = f"img_{int(time.time())}_{random_hex()}.{ext}" + image_path = UPLOADS_DIR / stored_name + uploaded.save(str(image_path)) + + snippet = f"" + return read_partial("upload_result.html", { + "{{ filename }}": str(escape(stored_name)), + "{{ markdown_snippet }}": str(escape(snippet)), + "{{ preview_url }}": f"/uploads/{stored_name}", + }) + + @app.route("/uploads/<path:filename>") + def serve_upload(filename): + if not is_safe_relative_path(filename): + abort(400) + return send_from_directory(str(UPLOADS_DIR), filename, conditional=True) + + @app.route("/download/<path:filename>") + def download(filename): + if not is_safe_relative_path(filename): + abort(400) + return send_from_directory( + str(GENERATED_DIR), + filename, + as_attachment=True, + download_name=filename, + conditional=True, + ) + + return app + + +app = create_app() + + +if __name__ == "__main__": + host = os.getenv("HOST", DEFAULT_HOST) + port = int(os.getenv("PORT", str(DEFAULT_PORT))) + print(f" {APP_NAME} listening on http://{host}:{port}") + app.run(host=host, port=port, debug=False) diff --git a/src/latex/template.tex b/src/latex/template.tex deleted file mode 100644 index 8f9aacc..0000000 --- a/src/latex/template.tex +++ /dev/null @@ -1,88 +0,0 @@ -\documentclass[11pt]{article} - -\usepackage{fontspec} -\newfontfamily{\emojifont}{Noto Color Emoji}[Renderer=HarfBuzz] -\directlua{ - luaotfload.add_fallback("emojifallback", { - "Noto Color Emoji:mode=harf;" - }) -} -\setmainfont{$mainfont$}[RawFeature={fallback=emojifallback}] -\setmonofont{Latin Modern Mono} - -$if(paperwidth)$ -\usepackage[paperwidth=$paperwidth$,paperheight=$paperheight$,margin=$margin$]{geometry} -$else$ -\usepackage[paper=$papersize$,margin=$margin$]{geometry} -$endif$ -\usepackage{microtype} -\usepackage{parskip} -\usepackage{setspace} -\setstretch{$linespacing$} -\usepackage{xcolor} -\usepackage{graphicx} -\usepackage{float} -\usepackage{booktabs} -\usepackage{longtable} -\usepackage{array} -\usepackage{calc} -\usepackage{etoolbox} -\usepackage{fancyvrb} -\usepackage{fvextra} -\DefineVerbatimEnvironment{Highlighting}{Verbatim}{breaklines,commandchars=\\\{\}} -\usepackage{hyperref} -\hypersetup{ - colorlinks=true, - linkcolor=black, - urlcolor=blue, - citecolor=black, - pdfauthor={}, - pdftitle={likha-pdf} -} -\urlstyle{same} - -\setlength{\emergencystretch}{3em} -\setcounter{secnumdepth}{0} -\setkeys{Gin}{width=\linewidth,keepaspectratio} -\makeatletter -\newsavebox\pandoc@box -\newcommand*\pandocbounded[1]{% - \sbox\pandoc@box{#1}% - \Gscale@div\@tempa{\textheight}{\dimexpr\ht\pandoc@box+\dp\pandoc@box\relax}% - \Gscale@div\@tempb{\linewidth}{\wd\pandoc@box}% - \ifdim\@tempb\p@<\@tempa\p@ - \scalebox{\@tempb}{\usebox\pandoc@box}% - \else - \scalebox{\@tempa}{\usebox\pandoc@box}% - \fi -} -\makeatother -\let\origfigure\figure -\let\endorigfigure\endfigure -\renewenvironment{figure}[1][] { - \expandafter\origfigure\expandafter[H] -} { - \endorigfigure -} -\providecommand{\tightlist}{ - \setlength{\itemsep}{0pt}\setlength{\parskip}{0pt} -} - -$if(highlighting-macros)$ -$highlighting-macros$ -$endif$ - -$if(hidepages)$ -\pagestyle{empty} -$endif$ - -\begin{document} - -$if(title)$ -{\huge\bfseries $title$\par} -\vspace{1em} -$endif$ - -$body$ - -\end{document}
\ No newline at end of file diff --git a/src/templates/index.html b/src/templates/index.html index 30c4b93..a9501e9 100644 --- a/src/templates/index.html +++ b/src/templates/index.html @@ -22,7 +22,7 @@ <body> <main> <h1>likha-pdf</h1> - <p>simple markdown export with pandoc + lualatex.</p> + <p>simple markdown to pdf export.</p> <form id="convert-form" hx-post="/convert" hx-target="#result" hx-swap="innerHTML" hx-indicator="#loading"> <label for="markdown"> |