MietGraph

Creating a knowledge graph about the rent stuff. This is a fork of BauGraph

Create the graph

To create the graph there are two steps. First we have to crawl a few websites then we create the graph.

1. Clone the repository git clone https://git.informatik.uni-leipzig.de/information_extraction_group/bauprofessor_crawler.git
2. Change into the repository folder and install the necessary requirements with pip install -r requirements.txt
3. Change into the directory cd MietGraph/crawling and run the crawling with ./run.sh [output_folder]
4. Change to MietGraph/graphing, move the output from step 3 to a folder for input and run python MietGraph.py --input=[input_folder] --output=MietGraph.graphml to generate the graph. (python MietGraph.py -h for more options)
5. ???
6. Profit

The output formats ".gml", ".gexf" and ".graphml" can all be used with gephi.

Installation

You should also be able to install the package with pip install . from the repository's base directory. After that you can invoke the MietGraph class and generate the Graph in python environment.

>> from MietGraph import MietGraph
>> g = MietGraph("input/")
>> g.add_keyword_in_text()
>> g.add_crosslinks()
>> g.graph
<networkx.classes.digraph.DiGraph at 0x7f38243927c0>

You can also get a igraph representation with

g.igraph()

You can use this to manipulate the graph further. More Information coming up..

---

KnowledgeBase

import MietGraph
kb = MietGraph.KnowledgeBase("name of your knowledge base file")

# See a list of contained documents
docs = kb.list_documents()

# Get all neighbours of a document with all attributes
kb.get_neighbors(docs[0], True)

#See the content of the document
kb.get_attr(docs[0], "content")

# Show document names of similar documents
kb.most_similar(docs[0])

# you can also add new pdfs to the knowledgebase
kb.add("path_to_new_pdf.pdf")

# And save the modified graph
kb.save("save.graphml")