Add orig text print before unl graph

9ea8564d · David Rouquet · e3beb791 · 9ea8564d
Commit 9ea8564d authored Jun 17, 2020 by David Rouquet
--- a/scripts/unlizeXmlNbSample.ipynb
+++ b/scripts/unlizeXmlNbSample.ipynb
@@ -8,6 +8,7 @@
   "source": [
    "import tempfile\n",
    "import os\n",
+    "import re\n",
    "from subprocess import Popen, PIPE, STDOUT\n",
    "from IPython.core.display import SVG"
   ]
@@ -73,6 +74,8 @@
    "    svgArray = [x+sep for x in protoSvg.split(sep)]\n",
    "    svgArray.pop()\n",
    "    for svg in svgArray :\n",
+    "        text = re.search('\\{org.*\\n(.*)\\n.*org\\}',unldata).group(1)\n",
+    "        print(text)\n",
    "        display(SVG(svg))"
   ]
  }
@@ -94,6 +97,13 @@
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.3"
+  },
+  "widgets": {
+   "application/vnd.jupyter.widget-state+json": {
+    "state": {},
+    "version_major": 2,
+    "version_minor": 0
+   }
  }
 },
 "nbformat": 4,

 %% Cell type:code id: tags:

 ``` python
 import tempfile
 import os
+import re
 from subprocess import Popen, PIPE, STDOUT
 from IPython.core.display import SVG
 ```

 %% Cell type:code id: tags:

 ``` python
 def unl2dot(text, path):
    with tempfile.NamedTemporaryFile() as temp:
        out_name = os.path.basename(temp.name)
        out_dir = os.path.dirname(temp.name)

    with tempfile.NamedTemporaryFile(mode="w") as in_file:
        # Remove CRLF and flush output to avoid java errors
        in_file.write(text.replace("\r\n", "\n"))
        in_file.flush()

        # Run java parser
        cmd = ['java', '-jar', path,
               '--input-file', in_file.name,
               '--output-Dir', out_dir, '--output-file', out_name,
               '--output-type', 'dot']

        with Popen(cmd, stdout=PIPE, stderr=STDOUT) as p:
            p.wait()
            p.stdout.flush()
            if p.returncode != 0:
                print("Error in unl2rdf: \n\n"+p.stdout.read().decode())
                print('UNL;')
                print(text)

    # generate dot output
    fname = '{}/{}.dot'.format(out_dir, out_name)
    cmd = ['dot', '-Tsvg', fname]
    with Popen(cmd, stdout=PIPE, stderr=PIPE) as p:
        p.wait()
        if p.returncode != 0:
            print("Error creating svg: \n\n"+p.stderr.read().decode())
            print('UNL:')
            print(text)
            try:
                with open(fname) as f:
                    print('DOT:')
                    print(f.read())
            except FileNotFoundError:
                pass
        else:
            svg = p.stdout.read().decode()
            os.remove(fname)
            return svg
    return ""


 def displayUnl(unldata) :
 # We generate protoSVG because whent there are several sentences,
 # a string composed of several concatenated SVG is produced (not a valid SVG).
 # We must then split the string to obtain several valid SVG to display.
    protoSvg = unl2dot(unldata, "unl2rdf-app-1.0-SNAPSHOT-jar-with-dependencies.jar")
    sep = "</svg>\n"
    svgArray = [x+sep for x in protoSvg.split(sep)]
    svgArray.pop()
    for svg in svgArray :
+        text = re.search('\{org.*\n(.*)\n.*org\}',unldata).group(1)
+        print(text)
        display(SVG(svg))
 ```