def run(message): tika_client = TikaApp(file_jar="./tika-app/tika-app-1.21.jar") tika_result = tika_client.extract_only_content(message["path"]) if (tika_result != None): processing_dir = "./data/processing/" identifier = str(uuid.uuid4()) workfile = processing_dir + identifier with open(workfile, 'wb') as f: f.write(tika_result.encode('UTF-8')) new_message = { "identifier": identifier, "parent": message["identifier"], "path": workfile, "filename": "pdf.txt", "filetype": "unknown", "history": [], "metadata": {}, "original_file": False } sendEvent(new_message) extract_images_pymupdf(message) sendEvent(message)
def main(): args = get_args() tika = TikaApp(args.jar or os.environ.get("TIKA_APP_JAR", None)) parameters = { "path": args.file, "payload": args.payload, "objectInput": sys.stdin if args.stdin else None } try: if args.detect: print(tika.detect_content_type(**parameters)) if args.text: print(tika.extract_only_content(**parameters)) if args.language: print(tika.detect_language(**parameters)) if args.all: parameters["pretty_print"] = True print(tika.extract_all_content(**parameters)) if args.metadata: parameters["pretty_print"] = True print(tika.extract_only_metadata(**parameters)) except IOError: pass
def main(): args = get_args() command_line = dict() if args.jar: command_line = {"TIKA_APP_JAR": args.jar} defaults = {"TIKA_APP_JAR": "/opt/tika/tika-app-1.15.jar"} options = ChainMap(command_line, os.environ, defaults) tika = TikaApp(options['TIKA_APP_JAR']) try: if args.file: f = args.file if args.detect: print(tika.detect_content_type(path=f)) if args.text: print(tika.extract_only_content(path=f)) if args.language: print(tika.detect_language(path=f)) if args.all: print(tika.extract_all_content(path=f, pretty_print=True)) elif args.payload: p = args.payload if args.detect: print(tika.detect_content_type(payload=p)) if args.text: print(tika.extract_only_content(payload=p)) if args.language: print(tika.detect_language(payload=p)) if args.all: print(tika.extract_all_content(payload=p, pretty_print=True)) except IOError: pass
def run(message): # Text tika_client = TikaApp(file_jar="./tika-app/tika-app-1.21.jar") tika_result = tika_client.extract_only_content(message["path"]) if (tika_result != None): processing_dir = "./data/processing/" identifier = str(uuid.uuid4()) workfile = processing_dir + identifier with open(workfile, 'wb') as f: f.write(tika_result.encode('UTF-8')) new_message = { "identifier": identifier, "parent": message["identifier"], "path": workfile, "filename" : "doc.txt", "filetype": "unknown", "history": [], "metadata": {}, "original_file": False } sendEvent(new_message) # Images with ZipFile(message["path"], 'r') as zipObj: processng_dir = "./data/processing" tmp_identifier = os.path.join(processng_dir, str(uuid.uuid4())) zipObj.extractall(tmp_identifier) for root, dirs, files in os.walk(tmp_identifier): for filename in files: if ".png" in filename or ".jpeg" in filename or ".jpg" in filename: new_identifier = str(uuid.uuid4()) processing_dest = os.path.join(processng_dir, new_identifier) new_message = { "identifier": new_identifier, "parent": message["identifier"], "path": processing_dest, "filename" : filename, "filetype": "unknown", "history": [], "metadata": {}, "original_file": False } shutil.move(os.path.join(root, filename), processing_dest) sendEvent(new_message) shutil.rmtree(tmp_identifier) sendEvent(message)
class TikaReader: # Iniciador de la clase. def __init__(self, file_process): # Cliente Tika que utiliza que carga el fichero jar cliente. self.tika_client = TikaApp(file_jar="tika-app-1.20.jar") self.file_process = file_process # Detector del tipo de contenido MIME. def detect_document_type(self): return self.tika_client.detect_content_type(self.file_process) # Detector de lenguaje utilizado en el documento. def detect_language(self): return self.tika_client.detect_language(self.file_process) # Extractor del contenido completo del documento. def extract_complete_info(self, value=False): return self.tika_client.extract_all_content(self.file_process, convert_to_obj=value) # Extractor de solo el contenido del documento. def extract_content_info(self): return self.tika_client.extract_only_content(self.file_process)
from tikapp import TikaApp tika_client = TikaApp( file_jar="/Users/yma2/Documents/_garage/python/cxm/tika/tika-app-1.20.jar") analyzeFile = "/Users/yma2/Downloads/Azure_Developer_Guide_eBook_ja-JP.pdf" print(tika_client.detect_content_type(analyzeFile)) print(tika_client.detect_language(analyzeFile)) print(tika_client.extract_only_content(analyzeFile)) print(tika_client.extract_only_metadata(analyzeFile))
def tika_extract_only_content(memory=None): tika_client = TikaApp(file_jar=TIKA_APP_JAR, memory_allocation=memory) output = tika_client.extract_only_content(path=test_zip) return output
def convert_Tika(self,fname): tika_client = TikaApp(file_jar=os.getcwd()+'/tika-app-1.20.jar') return tika_client.extract_only_content(fname)
def extract_text(filename): jar_path = os.path.abspath(os.path.join("lib", "tika-app-1.28.jar")) tika_client = TikaApp(file_jar=jar_path) parsed = tika_client.extract_only_content(filename) return parsed
class FileBeats: # # 建立文件索引 #只有如下扩展名的文件才会被索引文件内容 export_content_exts = ('.md', '.html', '.htm', '.txt', '.ppt', '.pptx', '.key', '.pdf', ".pages", ".doc", ".docx", '.py', '.java') def __init__(self, index, es_host="localhost:9200", file_jar='/Users/laofeng/es_home/tika-app-1.23.jar', index_content=False, create_index=True, force_renew_index=False, schema_file=None): self.tika_app = TikaApp(file_jar) self.es = Elasticsearch(es_host) self.index = index self.index_content = index_content #查询索引是否存在 index_exist = self.es.indices.exists(index) #if not index_exist and not create_index: if not index_exist and create_index and schema_file: with open(schema_file, 'r', encoding='utf-8') as f: schema = json.load(f) self.es.indices.create(index, schema) #如果已经存在,且强制renew,先删除后,再建立 if index_exist and force_renew_index: self.es.indices.delete(index) with open(schema_file, 'r', encoding='utf-8') as f: schema = json.load(f) self.es.indices.create(index, schema) # 格式化时间,参数是秒和时间格式 @staticmethod def second2date(second, style="%Y-%m-%d %H:%M:%S"): time_array = time.localtime(second) date_str = time.strftime(style, time_array) return date_str def export_file_tags(self, abs_path): tags = {"path": abs_path} (basename, ext) = os.path.splitext(abs_path) tags['ext'] = ext.lstrip('.') # 去掉了后缀的点 tags['name'] = os.path.basename(abs_path) size = os.path.getsize(abs_path) tags['size'] = size # 过滤调太大或者太小的文件 # 文件太大,二进制文件类型不导出content if self.index_content and ext.lower() in FileBeats.export_content_exts: try: r = self.tika_app.extract_only_content( path=abs_path, payload="base64_payload") if r: tags['content'] = r except Exception as e: traceback.print_exc() return tags # 索引文档 def index_doc(self, tags, _type='_doc'): # index 相当于表名, body被索引的文本(分词) tags['timestamp'] = datetime.now() # 使用文件全路径做为id res = self.es.index(index=self.index, doc_type=_type, body=tags, id=tags['path']) def index_docs(self, docs): for doc in docs: self.index_doc(doc) # 处理一个文件,先导出tags,然后索引文档 def process_file(self, f): tags = self.export_file_tags(f) self.index_doc(tags) def beats_more(self, folders, asynchronous=True): print("开始索引文件", FileBeats.second2date(time.time())) for folder in folders: self.start_beats(folder) print("索引文件结束", FileBeats.second2date(time.time())) def start_beats(self, source_dir='/Volumes/portable/sync/', asynchronous=True): print("开始索引文件", source_dir, FileBeats.second2date(time.time())) # 遍历文件 greenlets = list() #index_tasks = list() for folder, dirs, files in os.walk(source_dir, topdown=False): # 过滤掉一些文件夹 if '@' in folder or '.svn' in folder or folder.endswith( '.app') or "迅雷" in folder: print('忽略目录', folder) continue for f in files: if f.startswith("."): continue abs_path = os.path.join(folder, f) try: # process_file(abs_path) if asynchronous: greenlets.append( gevent.spawn(self.export_file_tags, abs_path)) else: tags = self.export_file_tags(abs_path) self.index_doc(tags) # 任务达到500个,执行一次 if len(greenlets) >= 5: gevent.joinall(greenlets) self.index_docs([g.value for g in greenlets]) # 使用并发es会出现一个read timeout或者是socket的错误 # index_tasks.append(gevent.spawn(index_docs,[g.value for g in greenlets])) greenlets.clear() # if len(index_tasks) > 50: # gevent.joinall(index_tasks) # index_tasks.clear() except Exception as e: traceback.print_exc() # 清理不足5个文件的情况 if asynchronous: gevent.joinall(greenlets) self.index_docs([g.value for g in greenlets]) # index_tasks.append(gevent.spawn(index_docs, [g.value for g in greenlets])) # gevent.joinall(index_tasks) print("索引文件结束", source_dir, FileBeats.second2date(time.time()))
def analyse_pdf_archive(pdf_csv, keyword_csv, tika_file_jar, outfile_name): # PDF for each day saved as welt_mmdd. File names listed in csv file 'welt_pdf'. Create list of PDF names. with open(pdf_csv, 'r', encoding = 'utf-8-sig') as f: reader = csv.reader(f) pdf_names = list(reader) pdf_names = list(itertools.chain(*pdf_names)) # acts as main data frame to contain individual data frames pdf_list = [] for name in pdf_names: pdf_list.append(name + '.pdf') # Create list of keywords from 'keyword_stems.csv' with open(keyword_csv, 'r', encoding = 'utf-8-sig') as f: reader = csv.reader(f) keywords = list(reader) keywords = list(itertools.chain(*keywords)) tika_client = TikaApp(file_jar=tika_file_jar) a = 1 #set counter to 1 keyword_counter = [] for pdf in pdf_list: rawText = tika_client.extract_only_content(pdf) print("pdf {0} extracted".format(a)) rawList = rawText.split( ) rawList_nopunct = [word.translate(str.maketrans('', '', string.punctuation)) for word in rawList] counts = Counter(rawList_nopunct) list_words = counts.most_common() keyword_hits_list = [] for x in range(0, len(list_words)): temp = list(list_words[x]) # convert from tuple to list temp[1] = str(temp[1]) # change number (at index 1) into string n_temp = [(unicodedata.normalize('NFKD', word).encode('ASCII', 'ignore')).lower().decode() for word in temp] #normalised umlauts in data #check word (at index 0) against list of keywords, add new column = 1 if match, = 0 otherwise. hits = 0 for i in range(0, len(keywords)): if keywords[i] in n_temp[0]: hits = hits + 1 if hits != 0: n_temp.append(1) else: n_temp.append(0) keyword_hits = int(n_temp[1])*int(n_temp[2]) keyword_hits_list.append(keyword_hits) keyword_counts = sum(keyword_hits_list) keyword_counter.append(keyword_counts) print("day {0} complete".format(a)) if list_words != []: a = a + 1 else: break df = pd.DataFrame({"id": pdf_names, "keywords": keyword_counter}) df.to_csv(outfile_name, index=False)