Update Dockerfile

Dockerfile

@@ -1,23 +1,151 @@
-FROM lfoppiano/grobid-evaluation:latest
-USER root
-RUN mkdir -m 777 -p /opt/grobid/grobid-home/tmp
-RUN mkdir -m 777 -p /opt/grobid/logs
-RUN chmod -R 777 /opt/grobid
-RUN chmod -R uog+rw /data/db
+import http.server
+import socketserver
+import subprocess
+import threading
+import queue
+import io
+import select
+import re
+import os
+import urllib.request
+import urllib.parse
+import urllib.error
 
-# CRF 
-COPY --chown=lfoppiano grobid.yaml /opt/grobid/grobid-home/config/grobid.yaml
+PORT = 8080
+command_executed = False
+command_output = queue.Queue()
+command_queue = queue.Queue()
 
-# official evaluation configuration 
-COPY --chown=lfoppiano grobid-full.yml /opt/grobid/grobid-home/config/grobid.yaml
-RUN ls -R /opt/grobid/evaluation/
 
-WORKDIR /opt/grobid
-RUN git lfs install && git clone --depth 1 https://huggingface.co/datasets/sciencialab/grobid-evaluation evaluation
-RUN chmod -R uog+rw /opt/grobid/evaluation
+def sanitize_output(output):
+    # Remove any non-printable characters except for carriage return
+    return re.sub(r'[^\x20-\x7E\r]+', '', output)
 
-COPY --chown=lfoppiano service.py /opt/grobid/service.py
 
-WORKDIR /opt/grobid 
+def run_command(command):
+    print("Running command ", command)
+    process = subprocess.Popen(
+        command,
+        stdout=subprocess.PIPE,
+        stderr=subprocess.PIPE,
+        text=True,
+        bufsize=1  # Line-buffered
+    )
+    while True:
+        reads = [process.stdout.fileno(), process.stderr.fileno()]
+        ret = select.select(reads, [], [], 0.1)  # Timeout to periodically flush
+        for fd in ret[0]:
+            if fd == process.stdout.fileno():
+                output = process.stdout.readline()
+                if output:
+                    command_output.put(sanitize_output(output) + "\n")
+            if fd == process.stderr.fileno():
+                error = process.stderr.readline()
+                if error:
+                    command_output.put(sanitize_output(error) + "\n")
+        if process.poll() is not None:
+            break
+        process.stdout.flush()
+        process.stderr.flush()
+    process.stdout.close()
+    process.stderr.close()
+    process.wait()
+    command_output.put("Leaving")
 
-CMD ["/bin/bash", "-c", "python service.py"]
+
+def command_worker():
+    print("Starting worker command")
+    while True:
+        command = command_queue.get()
+        if command is None:
+            break
+        run_command(command)
+        command_queue.task_done()
+
+
+def start_commands():
+    commands = [
+        ["echo", "## standard process"],
+        ['./gradlew', 'jatsEval', '-Pp2t=/opt/grobid/evaluation/PMC_sample_1943', '-Prun=1', '-PfileRatio=1'],
+        ["cp", "/opt/grobid/grobid-home/tmp/report.md", "/opt/grobid/grobid-home/tmp/report-pmc_sample_1943.md"],
+        ['./gradlew', 'jatsEval', '-Pp2t=/opt/grobid/evaluation/PLOS_1000', '-Prun=1', '-PfileRatio=1'],
+        ["cp", "/opt/grobid/grobid-home/tmp/report.md", "/opt/grobid/grobid-home/tmp/report-plos_1000.md"],
+        ['./gradlew', 'jatsEval', '-Pp2t=/opt/grobid/evaluation/biorxiv-10k-test-2000', '-Prun=1', '-PfileRatio=1'],
+        ["cp", "/opt/grobid/grobid-home/tmp/report.md", "/opt/grobid/grobid-home/tmp/report-biorxiv-10k-test-2000.md"],
+        ['./gradlew', 'jatsEval', '-Pp2t=/opt/grobid/evaluation/eLife_984', '-Prun=1', '-PfileRatio=1'],
+        ["cp", "/opt/grobid/grobid-home/tmp/report.md", "/opt/grobid/grobid-home/tmp/report-elife_984.md"],
+        ["echo", "## article/light"],
+        ['./gradlew', 'jatsEval', '-Pp2t=/opt/grobid/evaluation/PMC_sample_1943', '-Prun=1', '-PfileRatio=1', '-Pflavor=article/light'],
+        ["cp", "/opt/grobid/grobid-home/tmp/report.md", "/opt/grobid/grobid-home/tmp/report-pmc_sample_1943-article_light.md"],
+        ['./gradlew', 'jatsEval', '-Pp2t=/opt/grobid/evaluation/PLOS_1000', '-Prun=1', '-PfileRatio=1', '-Pflavor=article/light'],
+        ["cp", "/opt/grobid/grobid-home/tmp/report.md", "/opt/grobid/grobid-home/tmp/report-plos_1000-article_light.md"],
+        ['./gradlew', 'jatsEval', '-Pp2t=/opt/grobid/evaluation/biorxiv-10k-test-2000', '-Prun=1', '-PfileRatio=1', '-Pflavor=article/light'],
+        ["cp", "/opt/grobid/grobid-home/tmp/report.md", "/opt/grobid/grobid-home/tmp/report-biorxiv-10k-test-2000-article_light.md"],
+        ['./gradlew', 'jatsEval', '-Pp2t=/opt/grobid/evaluation/eLife_984', '-Prun=1', '-PfileRatio=1', '-Pflavor=article/light'],
+        ["cp", "/opt/grobid/grobid-home/tmp/report.md", "/opt/grobid/grobid-home/tmp/report-eLife_984-article_light.md"],
+        ["echo", "## article/light-ref"],
+        ['./gradlew', 'jatsEval', '-Pp2t=/opt/grobid/evaluation/PMC_sample_1943', '-Prun=1', '-PfileRatio=1', '-Pflavor=article/light-ref'],
+        ["cp", "/opt/grobid/grobid-home/tmp/report.md", "/opt/grobid/grobid-home/tmp/report-pmc_sample_1943-article_light_ref.md"],
+        ['./gradlew', 'jatsEval', '-Pp2t=/opt/grobid/evaluation/PLOS_1000', '-Prun=1', '-PfileRatio=1', '-Pflavor=article/light-ref'],
+        ["cp", "/opt/grobid/grobid-home/tmp/report.md", "/opt/grobid/grobid-home/tmp/report-plos_1000-article_light_ref.md"],
+        ['./gradlew', 'jatsEval', '-Pp2t=/opt/grobid/evaluation/biorxiv-10k-test-2000', '-Prun=1', '-PfileRatio=1', '-Pflavor=article/light-ref'],
+        ["cp", "/opt/grobid/grobid-home/tmp/report.md", "/opt/grobid/grobid-home/tmp/report-biorxiv-10k-test-2000-article_light_ref.md"],
+        ['./gradlew', 'jatsEval', '-Pp2t=/opt/grobid/evaluation/eLife_984', '-Prun=1', '-PfileRatio=1', '-Pflavor=article/light-ref'],
+        ["cp", "/opt/grobid/grobid-home/tmp/report.md", "/opt/grobid/grobid-home/tmp/report-eLife_984-article_light_ref.md"],
+    ]
+    for command in commands:
+        command_queue.put(command)
+    threading.Thread(target=command_worker).start()
+
+
+class Handler(http.server.SimpleHTTPRequestHandler):
+    def do_GET(self):
+        global command_executed
+
+        # Check if the request is for downloading a file from URL
+        if self.path.startswith('/fetch?filename='):
+            try:
+                # Parse the URL from the query string
+                filename = urllib.parse.unquote(self.path.split('=', 1)[1])
+
+                # Construct the full file path
+                file_path = os.path.join('/opt/grobid/grobid-home/tmp/', filename)
+
+                # Check if the file exists
+                if not os.path.isfile(file_path):
+                    self.send_error(404, f"File '{filename}' not found.")
+                    return
+
+                # Read the file content
+                with open(file_path, 'r') as file:
+                    content = file.read()
+
+                response = content
+
+            except urllib.error.URLError as e:
+                self.send_error(500, f"Failed to download file: {str(e)}")
+            except Exception as e:
+                self.send_error(500, f"An error occurred: {str(e)}")
+
+        elif not command_executed:
+            command_executed = True
+            threading.Thread(target=start_commands).start()
+            response = "Starting evaluation."
+
+        else:
+            response = io.StringIO()
+            response.write("Command output:\n")
+            for item in list(command_output.queue):
+                response.write(item)
+
+            response = response.getvalue()
+
+        self.send_response(200)
+        self.send_header("Content-type", "text/plain")
+        self.end_headers()
+        self.wfile.write(response.encode())
+
+
+with socketserver.TCPServer(("", PORT), Handler) as httpd:
+    print(f"Serving on port {PORT}")
+    httpd.serve_forever()