Atualiza Apache Tika e corrige build corrompido

Versões mais novas do Tika exigem o "Accept: text/plain" para
retornar apenas o conteúdo textual, pois o padrão é retornar HTML.

data_extraction/text_extraction.py

@@ -25,7 +25,10 @@ def _try_extract_text(self, filepath: str) -> str:
         if self.is_txt(filepath):
             return self._return_file_content(filepath)
         with open(filepath, "rb") as file:
-            headers = {"Content-Type": self._get_file_type(filepath)}
+            headers = {
+                "Content-Type": self._get_file_type(filepath),
+                "Accept": "text/plain",
+            }
             response = requests.put(f"{self._url}/tika", data=file, headers=headers)
             response.encoding = "UTF-8"
             return response.text

scripts/Dockerfile_apache_tika

@@ -6,7 +6,7 @@ RUN adduser --system gazette && \
 	apt-get clean
 
 # install Apache Tika
-RUN curl -o /tika-server.jar http://archive.apache.org/dist/tika/tika-server-1.24.1.jar && \
+RUN curl -o /tika-server.jar https://archive.apache.org/dist/tika/2.9.2/tika-server-standard-2.9.2.jar && \
 	chmod 755 /tika-server.jar
 
 USER gazette