Merge remote-tracking branch 'gwdg/main'

2012244c · Christian Boulanger · 9d74d363 · 48cb0887 · 2012244c · 2012244c
Commit 2012244c authored 11 months ago by Christian Boulanger
--- a/prodigy/anystyle-to-prodigy.ipynb
+++ b/prodigy/anystyle-to-prodigy.ipynb
+{
+ "cells": [
+  {
+   "metadata": {},
+   "cell_type": "markdown",
+   "source": [
+    "# Convert AnyStyle training files to the Prodigy format\n",
+    "\n",
+    "Converter functions written by GPT-4, see https://chat.openai.com/share/3f42ae1d-3066-4563-944d-3139b5e1e867"
+   ],
+   "id": "cb631197c03a0768"
+  },
+  {
+   "metadata": {},
+   "cell_type": "markdown",
+   "source": "## Parser files",
+   "id": "83bfc356ce72fe51"
+  },
+  {
+   "cell_type": "code",
+   "id": "initial_id",
+   "metadata": {
+    "collapsed": true,
+    "ExecuteTime": {
+     "end_time": "2024-04-12T05:27:47.466154Z",
+     "start_time": "2024-04-12T05:27:47.459668Z"
+    }
+   },
+   "source": [
+    "import xml.etree.ElementTree as ET\n",
+    "import json\n",
+    "\n",
+    "def xml_to_jsonl(input_xml_path, output_jsonl_path):\n",
+    "    tree = ET.parse(input_xml_path)\n",
+    "    root = tree.getroot()\n",
+    "\n",
+    "    with open(output_jsonl_path, 'w', encoding='utf-8') as f:\n",
+    "        for sequence in root.findall('sequence'):\n",
+    "            text = ''\n",
+    "            spans = []\n",
+    "            current_pos = 0\n",
+    "\n",
+    "            for element in sequence:\n",
+    "                content = (element.text or '').strip() + ' '  # Append a space for separation\n",
+    "                if content:\n",
+    "                    start = current_pos\n",
+    "                    end = start + len(content) - 1  # Subtract 1 to account for the space added\n",
+    "                    spans.append({'start': start, 'end': end, 'label': element.tag})\n",
+    "                    text += content\n",
+    "                    current_pos = end + 1  # Move current position to right after the last character\n",
+    "\n",
+    "            # Store the reconstructed sequence and its spans in a dictionary\n",
+    "            entry = {'text': text.strip(), 'spans': spans}\n",
+    "\n",
+    "            # Write the JSON formatted string to the file, ending with a newline\n",
+    "            f.write(json.dumps(entry) + '\\n')\n"
+   ],
+   "outputs": [],
+   "execution_count": 1
+  },
+  {
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2024-04-12T06:23:46.935341Z",
+     "start_time": "2024-04-12T06:23:46.920522Z"
+    }
+   },
+   "cell_type": "code",
+   "source": [
+    "from pathlib import Path\n",
+    "for filename in [file.stem for file in Path('in').glob('*.xml')]:\n",
+    "    xml_to_jsonl(f'in/{filename}.xml', f'out/{filename}-parser.jsonl')"
+   ],
+   "id": "fa9f2c194697a6b7",
+   "outputs": [],
+   "execution_count": 8
+  },
+  {
+   "metadata": {},
+   "cell_type": "markdown",
+   "source": "## Finder files",
+   "id": "2988615a1be96bb"
+  },
+  {
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2024-04-12T06:54:04.924138Z",
+     "start_time": "2024-04-12T06:54:04.913099Z"
+    }
+   },
+   "cell_type": "code",
+   "source": [
+    "import json\n",
+    "\n",
+    "def markup_to_jsonl(input_path, output_jsonl_path, break_tags_before=None, break_tags_after=None):\n",
+    "    if break_tags_before is None:\n",
+    "        break_tags_before = []\n",
+    "    if break_tags_after is None:\n",
+    "        break_tags_after = []\n",
+    "\n",
+    "    # Helper function to write blocks to JSONL, now takes a file handle\n",
+    "    def write_block(full_text, spans, file_handle):\n",
+    "        if full_text.strip():  # Ensure there's content to write\n",
+    "            result = {\n",
+    "                'text': full_text.rstrip(),  # Remove trailing spaces\n",
+    "                'spans': spans\n",
+    "            }\n",
+    "            file_handle.write(json.dumps(result) + '\\n')\n",
+    "\n",
+    "    # Open the output file once at the beginning\n",
+    "    with open(output_jsonl_path, 'w', encoding='utf-8') as f:\n",
+    "        with open(input_path, 'r', encoding='utf-8') as file:\n",
+    "            lines = file.readlines()\n",
+    "\n",
+    "        full_text = ''\n",
+    "        spans = []\n",
+    "        current_text = ''\n",
+    "        text_start_pos = 0\n",
+    "        current_tag = 'text'  # Default initial tag\n",
+    "\n",
+    "        # Iterate through lines\n",
+    "        for line in lines:\n",
+    "            parts = line.strip().split('|', 1)\n",
+    "            if len(parts) < 2:\n",
+    "                continue\n",
+    "\n",
+    "            tag = parts[0].strip()\n",
+    "            content = parts[1].strip()\n",
+    "\n",
+    "            if tag in break_tags_before:\n",
+    "                # Finish the current text block before starting new due to break tag\n",
+    "                if current_text:\n",
+    "                    full_text += current_text\n",
+    "                    spans.append({\n",
+    "                        'start': text_start_pos,\n",
+    "                        'end': text_start_pos + len(current_text),\n",
+    "                        'label': current_tag\n",
+    "                    })\n",
+    "                    write_block(full_text, spans, f)\n",
+    "                    full_text, spans, current_text = '', [], ''\n",
+    "                current_tag = tag  # Update to the new tag\n",
+    "                text_start_pos = len(full_text)\n",
+    "\n",
+    "            # Append content with maintaining line break if the previous content was not empty\n",
+    "            current_text += content + '\\n' if current_text else content\n",
+    "\n",
+    "            if tag:\n",
+    "                if current_text and current_text != content:  # Ensures not to add just processed content span\n",
+    "                    # Finish the current span\n",
+    "                    full_text += current_text\n",
+    "                    spans.append({\n",
+    "                        'start': text_start_pos,\n",
+    "                        'end': text_start_pos + len(current_text) - 1,  # -1 to not count the last \\n for span\n",
+    "                        'label': current_tag\n",
+    "                    })\n",
+    "                current_text = content + '\\n'  # Start new text content under the new tag\n",
+    "                current_tag = tag\n",
+    "                text_start_pos = len(full_text)\n",
+    "\n",
+    "            if tag in break_tags_after:\n",
+    "                # Finalize current text and write it due to break tag\n",
+    "                full_text += current_text\n",
+    "                spans.append({\n",
+    "                    'start': text_start_pos,\n",
+    "                    'end': text_start_pos + len(current_text) - 1,  # -1 as above\n",
+    "                    'label': current_tag\n",
+    "                })\n",
+    "                current_text = ''\n",
+    "                write_block(full_text, spans, f)\n",
+    "                full_text, spans = '', []\n",
+    "\n",
+    "        # Final write for any remaining text\n",
+    "        if current_text:\n",
+    "            full_text += current_text\n",
+    "            spans.append({\n",
+    "                'start': text_start_pos,\n",
+    "                'end': text_start_pos + len(current_text) - 1,  # -1 to adjust for the last newline character\n",
+    "                'label': current_tag\n",
+    "            })\n",
+    "        write_block(full_text, spans, f)\n",
+    "\n",
+    "# Example usage:\n",
+    "# markup_to_jsonl('input_text.txt', 'output.jsonl', ['header'], ['footer'])\n"
+   ],
+   "id": "7764c66e149abd4f",
+   "outputs": [],
+   "execution_count": 26
+  },
+  {
+   "metadata": {},
+   "cell_type": "markdown",
+   "source": "Since the finder files have no information on page breaks, document segmentation is done using the `meta` tag, which is a very simple heuristic and probably fails in many cases. ",
+   "id": "8923ed5011c1d65f"
+  },
+  {
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2024-04-12T06:54:05.879064Z",
+     "start_time": "2024-04-12T06:54:05.870910Z"
+    }
+   },
+   "cell_type": "code",
+   "source": [
+    "from pathlib import Path\n",
+    "for filename in [file.stem for file in Path('in').glob('*.xml')]:\n",
+    "    markup_to_jsonl(f'in/{filename}.ttx', f'out/{filename}-finder.jsonl', break_tags_before=['meta'])"
+   ],
+   "id": "2b0ca72bbf70c829",
+   "outputs": [],
+   "execution_count": 27
+  },
+  {
+   "metadata": {},
+   "cell_type": "code",
+   "outputs": [],
+   "execution_count": null,
+   "source": "",
+   "id": "46d7e63e53b441e3"
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 2
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython2",
+   "version": "2.7.6"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
+%% Cell type:markdown id:cb631197c03a0768 tags:
+# Convert AnyStyle training files to the Prodigy format
+Converter functions written by GPT-4, see https://chat.openai.com/share/3f42ae1d-3066-4563-944d-3139b5e1e867
+%% Cell type:markdown id:83bfc356ce72fe51 tags:
+## Parser files
+%% Cell type:code id:initial_id tags:
+``` python
+import xml.etree.ElementTree as ET
+import json
+def xml_to_jsonl(input_xml_path, output_jsonl_path):
+    tree = ET.parse(input_xml_path)
+    root = tree.getroot()
+    with open(output_jsonl_path, 'w', encoding='utf-8') as f:
+        for sequence in root.findall('sequence'):
+            text = ''
+            spans = []
+            current_pos = 0
+            for element in sequence:
+                content = (element.text or '').strip() + ' '  # Append a space for separation
+                if content:
+                    start = current_pos
+                    end = start + len(content) - 1  # Subtract 1 to account for the space added
+                    spans.append({'start': start, 'end': end, 'label': element.tag})
+                    text += content
+                    current_pos = end + 1  # Move current position to right after the last character
+            # Store the reconstructed sequence and its spans in a dictionary
+            entry = {'text': text.strip(), 'spans': spans}
+            # Write the JSON formatted string to the file, ending with a newline
+            f.write(json.dumps(entry) + '\n')
+```
+%% Cell type:code id:fa9f2c194697a6b7 tags:
+``` python
+from pathlib import Path
+for filename in [file.stem for file in Path('in').glob('*.xml')]:
+    xml_to_jsonl(f'in/{filename}.xml', f'out/{filename}-parser.jsonl')
+```
+%% Cell type:markdown id:2988615a1be96bb tags:
+## Finder files
+%% Cell type:code id:7764c66e149abd4f tags:
+``` python
+import json
+def markup_to_jsonl(input_path, output_jsonl_path, break_tags_before=None, break_tags_after=None):
+    if break_tags_before is None:
+        break_tags_before = []
+    if break_tags_after is None:
+        break_tags_after = []
+    # Helper function to write blocks to JSONL, now takes a file handle
+    def write_block(full_text, spans, file_handle):
+        if full_text.strip():  # Ensure there's content to write
+            result = {
+                'text': full_text.rstrip(),  # Remove trailing spaces
+                'spans': spans
+            }
+            file_handle.write(json.dumps(result) + '\n')
+    # Open the output file once at the beginning
+    with open(output_jsonl_path, 'w', encoding='utf-8') as f:
+        with open(input_path, 'r', encoding='utf-8') as file:
+            lines = file.readlines()
+        full_text = ''
+        spans = []
+        current_text = ''
+        text_start_pos = 0
+        current_tag = 'text'  # Default initial tag
+        # Iterate through lines
+        for line in lines:
+            parts = line.strip().split('|', 1)
+            if len(parts) < 2:
+                continue
+            tag = parts[0].strip()
+            content = parts[1].strip()
+            if tag in break_tags_before:
+                # Finish the current text block before starting new due to break tag
+                if current_text:
+                    full_text += current_text
+                    spans.append({
+                        'start': text_start_pos,
+                        'end': text_start_pos + len(current_text),
+                        'label': current_tag
+                    })
+                    write_block(full_text, spans, f)
+                    full_text, spans, current_text = '', [], ''
+                current_tag = tag  # Update to the new tag
+                text_start_pos = len(full_text)
+            # Append content with maintaining line break if the previous content was not empty
+            current_text += content + '\n' if current_text else content
+            if tag:
+                if current_text and current_text != content:  # Ensures not to add just processed content span
+                    # Finish the current span
+                    full_text += current_text
+                    spans.append({
+                        'start': text_start_pos,
+                        'end': text_start_pos + len(current_text) - 1,  # -1 to not count the last \n for span
+                        'label': current_tag
+                    })
+                current_text = content + '\n'  # Start new text content under the new tag
+                current_tag = tag
+                text_start_pos = len(full_text)
+            if tag in break_tags_after:
+                # Finalize current text and write it due to break tag
+                full_text += current_text
+                spans.append({
+                    'start': text_start_pos,
+                    'end': text_start_pos + len(current_text) - 1,  # -1 as above
+                    'label': current_tag
+                })
+                current_text = ''
+                write_block(full_text, spans, f)
+                full_text, spans = '', []
+        # Final write for any remaining text
+        if current_text:
+            full_text += current_text
+            spans.append({
+                'start': text_start_pos,
+                'end': text_start_pos + len(current_text) - 1,  # -1 to adjust for the last newline character
+                'label': current_tag
+            })
+        write_block(full_text, spans, f)
+# Example usage:
+# markup_to_jsonl('input_text.txt', 'output.jsonl', ['header'], ['footer'])
+```
+%% Cell type:markdown id:8923ed5011c1d65f tags:
+Since the finder files have no information on page breaks, document segmentation is done using the `meta` tag, which is a very simple heuristic and probably fails in many cases. 
+%% Cell type:code id:2b0ca72bbf70c829 tags:
+``` python
+from pathlib import Path
+for filename in [file.stem for file in Path('in').glob('*.xml')]:
+    markup_to_jsonl(f'in/{filename}.ttx', f'out/{filename}-finder.jsonl', break_tags_before=['meta'])
+```
+%% Cell type:code id:46d7e63e53b441e3 tags:
+``` python
+```
--- a/prodigy/in/10.1515_zfrs-1980-0103.ttx
+++ b/prodigy/in/10.1515_zfrs-1980-0103.ttx
--- a/prodigy/in/10.1515_zfrs-1980-0103.xml
+++ b/prodigy/in/10.1515_zfrs-1980-0103.xml
+<?xml version="1.0" encoding="utf-8"?>
+<dataset>
+    <sequence>
+        <citation-number>1</citation-number>
+        <author>Geiger</author>
+        <date>1964,</date>
+        <pages>insbesondere S. 65—83.</pages>
+    </sequence>
+    <sequence>
+        <citation-number>2</citation-number>
+        <signal>Vgl.</signal>
+        <author>Feest/Blankenburg,</author>
+        <date>1972.</date>
+        <signal>Die Konsequenz einer größeren Dunkelziffer bei den von der Polizei selbst entdeckten Straftaten entwickle ich ausführlicher in meinem Beitrag über </signal>
+        <title>,Nichtkriminalisierung als Struktur und Routine',</title>
+        <date>1976.</date>
+    </sequence>
+    <sequence>
+        <citation-number>3</citation-number>
+        <note>Angaben aus einer Befragung von Peter MacNaughton-Smith und Richard Rosellen zur</note>
+        <title>.Bereitschaft zur Anzeigeerstattung' </title>
+        <note>Manuskript</note>
+        <publisher> Max-Planck-Institut für Strafrecht,</publisher>
+        <location>Freiburg</location>
+        <date>1978.</date>
+        <signal>Der ausführliche Forschungsbericht von </signal>
+        <author>Richard  Rosellen</author>
+        <note> erscheint in Kürze unter dem Titel</note>
+        <title>.Private Verbrechenskontrolle — eine empirische Untersuchung zur Anzeigeerstattung',</title>
+        <location>Berlin,</location>
+        <date>voraussichtlich 1980.</date>
+    </sequence>
+    <sequence>
+        <citation-number>4</citation-number>
+        <signal>Vgl.</signal>
+        <author>Blankenburg/Sessar/Steffen,</author>
+        <date>1978,</date>
+        <pages>S. 66-85.</pages>
+    </sequence>
+    <sequence>
+        <citation-number>5</citation-number>
+        <author>Black</author>
+        <date>1973,</date>
+        <pages>S. 125 ff.</pages>
+    </sequence>
+    <sequence>
+        <citation-number>6</citation-number>
+        <signal>Zur höheren Wahrscheinlichkeit der Normierung von Verhalten in weniger komplexen Beziehungen vgl. die Konflikttheorie von</signal>
+        <author>Gessner</author>
+        <date>1976,</date>
+        <pages>insbesondere S. 170—183.</pages>
+    </sequence>
+    <sequence>
+        <citation-number>7</citation-number>
+        <signal>Zum Konzept der .Thematisierung von Recht' vgl.</signal>
+        <journal>Luhmann</journal>
+        <date>1980,</date>
+        <pages>S. 99—112.</pages>
+    </sequence>
+    <sequence>
+        <citation-number>8</citation-number>
+        <signal>Ausführlicher bei</signal>
+        <author>Gessner</author>
+        <date>1976</date>
+    </sequence>
+    <sequence>
+        <citation-number>9</citation-number>
+        <signal>Vgl.</signal>
+        <author>Schönholz</author>
+        <date>1980.</date>
+    </sequence>
+    <sequence>
+        <citation-number>10</citation-number>
+        <author>Blankenburg/Schönholz; Rogowski</author>
+        <date>1979,</date>
+        <pages>S. 64 ff.</pages>
+    </sequence>
+    <sequence>
+        <citation-number>11</citation-number>
+        <author>Hilden</author>
+        <date>1976,</date>
+        <pages>S. 64 ff.</pages>
+    </sequence>
+    <sequence>
+        <citation-number>12</citation-number>
+        <author>Koch</author>
+        <date>1975,</date>
+        <pages>S. 75,</pages>
+        <note>der allerdings nur streitige Urteile und Vergleiche untersucht hat. </note>
+    </sequence>
+    <sequence>
+        <ignore>Mobilisierung von Recht 41</ignore>
+    </sequence>
+    <sequence>
+        <citation-number>13</citation-number>
+        <signal>Für Angaben der Zählkartenstatistik für die Familiengerichte siehe:</signal>
+        <author>Statistisches Bundesamt Wiesbaden,</author>
+        <title>Fachserie 10 (Rechtspflege) Reihe 2.1, Tabelle 10,</title>
+        <location>Wiesbaden</location>
+        <date>1978.</date>
+    </sequence>
+    <sequence>
+        <citation-number>14</citation-number>
+        <author>Blankenburg/Schönholz; Rogowski</author>
+        <date>1979,</date>
+        <pages>S. 78 ff.</pages>
+    </sequence>
+    <sequence>
+        <citation-number>15</citation-number>
+        <note>Steinbach 1979 hat in einer Erhebung von 1144 Amtsgerichtsprozessen in der Bundesrepublik ohne Berlin in den Jahren 1974—1976 ein Verhältnis von 1 Räumungsklage je 1,2 Forderungsklagen, in Berlin allerdings von 1 Räumungsklage je 0,12 Forderungsklagen ermittelt. Im folgenden auch als GMD-Erhebung zitiert.</note>
+    </sequence>
+    <sequence>
+        <citation-number>16</citation-number>
+        <author>Johnson</author>
+        <date>1979</date>
+        <note> berichtet von dem erfolgreichen Versuch einer Unfall-Schadensregulierung ohne Berücksichtigung einer Schuldzuschreibung in Neuseeland (no fault accident compensation),</note>
+        <pages>S. 90 ff.</pages>
+    </sequence>
+    <sequence>
+        <citation-number>17</citation-number>
+        <location>Steinbach</location>
+        <date>1979,</date>
+        <pages>S. 66 ff.;</pages>
+        <author>Blankenburg/Blankenburg/Morasch</author>
+        <date>1972,</date>
+        <pages>S. 82 ff.</pages>
+    </sequence>
+    <sequence>
+        <citation-number>18</citation-number>
+        <title>Projektbericht ,Rechtshilfebedürfnisse sozial Schwacher*,</title>
+        <author>(Blankenburg/Gorges/Reifner; Ticmann). </author>
+        <note>Befragt  wurden im Januar bis März 1979 je eine Person in 835 Haushalten Westberlins. Veröffentlichung ist vorgesehen für Ende 1980. Quelle: Eigene Befragung in West-Berlin 1979, Zu falls auswähl aus allen Haushalten mit Deutschen. Mehrfachnennungen, % jeweils bezogen auf Personen, die im jeweiligen Bereich ein Problem nennen.</note>
+    </sequence>
+    <sequence>
+        <citation-number>19</citation-number>
+        <signal>Explizit bei</signal>
+        <author>Baumgärtei</author>
+        <date>1976,</date>
+        <pages>S. 113-128.</pages>
+    </sequence>
+    <sequence>
+        <citation-number>20</citation-number>
+        <note>Laut einer GMD-Erhebung aus der Zählkartenstatistik wurde in Baden-Württemberg 1978 in</note>
+    </sequence>
+    <sequence>
+        <citation-number>21</citation-number>
+        <signal>Projektbericht</signal>
+        <author>.Rechtsschutzversicherung* (Blankenburg; Fiedler),</author>
+        <title>Veröffentlichung voraussichtlich Mitte</title>
+        <date>1980.</date>
+    </sequence>
+    <sequence>
+        <citation-number>22</citation-number>
+        <signal>Vgl. auch</signal>
+        <author>Reifner/Gorges</author>
+        <date>1980.</date>
+    </sequence>
+    <sequence>
+        <citation-number>23</citation-number>
+        <author>Reifner</author>
+        <date>1979.</date>
+    </sequence>
+    <sequence>
+        <citation-number>24</citation-number>
+        <signal>Klassisch bei</signal>
+        <author>Carlin/Howard/Messinger</author>
+        <date>1967.</date>
+    </sequence>
+    <sequence>
+        <citation-number>25</citation-number>
+        <signal>Grundsätzlich vgl. hierzu den klassischen Beitrag von</signal>
+        <author>Galanter</author>
+        <date>1974,</date>
+        <pages>S. 95—160.</pages>
+        <note>Die grösseren Chancen von Firmen, insbesondere bei der großen Zahl von vorstreitigen Erledigungen zeigt </note>
+        <author>Sarat</author>
+        <date>1976,</date>
+        <pages>S. 339-375.</pages>
+    </sequence>
+    <sequence>
+        <citation-number>26</citation-number>
+        <author>Bender/Schumachcr</author>
+        <date>1980,</date>
+        <pages>S. 138.</pages>
+    </sequence>
+    <sequence>
+        <citation-number>2</citation-number>
+        <title>Eigene Erhebung,</title>
+        <location>Arbeitsgericht Berlin, Gerichtsregister</location>
+        <date>1976</date>
+    </sequence>
+    <sequence>
+        <citation-number>3</citation-number>
+        <author>Zählkarten-Statistik, Statist. Bundesamt:</author>
+        <title>Fachserie 10 (Rechtspflege), Reihe</title>
+        <volume>2, 1 (Zivilgerichte),</volume>
+        <date>1971 und 1978</date>
+    </sequence>
+    <sequence>
+        <citation-number>27</citation-number>
+        <location>Steinbach</location>
+        <date>1979,</date>
+        <pages>S. 96 ff,</pages>
+        <signal>vgl. auch</signal>
+        <author>Blankenburg/Blankenburg/Morasch</author>
+        <date>1972,</date>
+        <pages>S. 89, Fn. 17,</pages>
+    </sequence>
+    <sequence>
+        <citation-number>28</citation-number>
+        <author>Reifner</author>
+        <date>1978.</date>
+    </sequence>
+    <sequence>
+        <citation-number>29</citation-number>
+        <author>Blankenburg/Schönholz; Rogowski</author>
+        <date>1979,</date>
+        <pages>S. 102 ff.</pages>
+    </sequence>
+    <sequence>
+        <citation-number>30</citation-number>
+        <signal>Vgl. zum</signal>
+        <title>.Verrechtlichungs’begriff meinen Beitrag über: Recht als gradualisiertes Konzept</title>
+        <date>1980,</date>
+        <pages>S. 83—98.</pages>
+    </sequence>
+    <sequence>
+        <citation-number>31</citation-number>
+        <location>Steinbach</location>
+        <date>1979,</date>
+        <pages>S. 35;</pages>
+        <author>Bender/Schumacher</author>
+        <date>1980,</date>
+        <pages>S. 24 und S. 49.</pages>
+    </sequence>
+    <sequence>
+        <citation-number>32</citation-number>
+        <author>Wanner</author>
+        <date>1975,</date>
+        <pages>S. 293—306</pages>
+        <note>hebt dies deutlich hervor, ebenso Bender/Schumacher</note>
+        <date>1980, </date>
+        <pages>S. 72 ff.; </pages>
+        <signal>sowie</signal>
+        <author>Galanter</author>
+        <date>1974; </date>
+        <author>Sarat</author>
+        <date> 1976.</date>
+    </sequence>
+    <sequence>
+        <citation-number>33</citation-number>
+        <author>Blankenburg/Schönholz; Rogowski</author>
+        <date>1979,</date>
+        <pages>S. 78 ff.</pages>
+    </sequence>
+    <sequence>
+        <citation-number>34</citation-number>
+        <backref>Ebenda,</backref>
+        <pages>S. 138-186.</pages>
+    </sequence>
+    <sequence>
+        <citation-number>35</citation-number>
+        <author>Luhmann</author>
+        <date>1969.</date>
+    </sequence>
+    <sequence>
+        <citation-number>36</citation-number>
+        <signal>Für eine überaus positive Bewertung des .Vermeidens1 vgl.</signal>
+        <author>Felstiner und Danzig/Lowy</author>
+        <journal>in Law and Society Review,</journal>
+        <volume>vol. 9,</volume>
+        <date>1974/75.</date>
+    </sequence>
+    <sequence>
+        <author>Baumgärtei, Gottfried,</author>
+        <date>1976:</date>
+        <title>Gleicher Zugang zum Recht für alle.</title>
+        <location>Köln.</location>
+    </sequence>
+    <sequence>
+        <author>Bender, Rolf und Rolf Schumacher,</author>
+        <date>1980:</date>
+        <title>Erfolgsbarrieren vor Gericht. </title>
+        <location>Tübingen.</location>
+        <author>Black, Donald,</author>
+        <date>1973:</date>
+        <title>The Mobilization of Law,</title>
+        <journal>in: Journal of Legal Studies</journal>
+        <volume>2.</volume>
+    </sequence>
+    <sequence>
+        <author>Blankenburg, Erhard/Blankenburg, Viola/Morasch, Helmut,</author>
+        <date>1972:</date>
+        <title>Der lange Weg in die Berufung,</title>
+        <editor>in: Bender, Rolf (Hrsg.):</editor>
+        <container-title>Tatsachenforschung in der Justiz.</container-title>
+        <location>Tübingen.</location>
+    </sequence>
+    <sequence>
+        <author>Blankenburg, Erhard,</author>
+        <date>1976:</date>
+        <title>Nichtkriminalisierung als Struktur und Routine,</title>
+        <editor>in: Göppinger,</editor>
+        <title>Hans und Günter Kaiser: Kriminologie und Strafverfahren.</title>
+        <location>Stuttgart.</location>
+    </sequence>
+    <sequence>
+        <author>Blankenburg, Erhard/Sessar, Klaus/Steffen, Wiebke,</author>
+        <date>1978:</date>
+        <title>Die Staatsanwaltschaft im Prozeß strafrechtlicher Sozialkontrolle.</title>
+        <location>Berlin.</location>
+    </sequence>
+    <sequence>
+        <author>Blankenburg, Erhard; Schönholz, Siegfried, unter Mitarbeit von Ralf Rogowski,</author>
+        <date>1979:</date>
+        <title>Zur Soziologie des Arbeitsgerichtsverfahrens.</title>
+        <location>Neuwied und Darmstadt.</location>
+    </sequence>
+    <sequence>
+        <author>Blankenburg, Erhard,</author>
+        <date>1980:</date>
+        <title>Recht als gradualisiertes Konzept,</title>
+        <editor>in: Blankenburg, Erhard; Klausa, Ekkehard und Hubert Rottleuthner (Hrsg.):</editor>
+        <container-title>Alternative Rechtsformen und Alternativen zum Recht,</container-title>
+        <collection-title>Jahrbuch für Rechtssoziologie und Rechtstheorie</collection-title>
+        <volume>Bd. 6. Opladen.</volume>
+    </sequence>
+    <sequence>
+        <author>Carlin, Jerome-, Jan Howard und Sheldon Messinger,</author>
+        <date>1967:</date>
+        <title>Civil Justice and the Poor.</title>
+        <location>New York. Danzig, </location>
+        <author>Richard /Lowy, Michael,</author>
+        <date>1974/75;</date>
+        <title>Everday Disputes and Mediation in the United States: A Reply to Professor Felstiner,</title>
+        <journal>Law and Society Review,</journal>
+        <volume>vol. 9,</volume>
+        <pages>pp. 675—694.</pages>
+    </sequence>
+    <sequence>
+        <author>Feest, Johannes/Blankenburg, Erhard,</author>
+        <date>1972:</date>
+        <title>Die Definitionsmacht der Polizei.</title>
+        <location>Düsseldorf.</location>
+    </sequence>
+    <sequence>
+        <author>Felstiner, William L. F.,</author>
+        <date>1974/75:</date>
+        <title>Influences of Social Organization on Dispute processing,</title>
+        <journal>in: Law and Society Review,</journal>
+        <volume>vol. 9,</volume>
+        <pages>pp. 63—94.</pages>
+    </sequence>
+    <sequence>
+        <author>Felstiner, William L. F.,</author>
+        <date>1974/75:</date>
+        <title>Avoidance as Dispute Processing: An Elaboration,</title>
+        <journal>in: Law and Society Review,</journal>
+        <volume>vol. 9,</volume>
+        <pages>pp. 695-706.</pages>
+    </sequence>
+    <sequence>
+        <author>Galanter, Marc,</author>
+        <date>1974:</date>
+        <title>Why the ,Haves* Come out Ahead: Speculations on the Limits of Legal Change,</title>
+        <journal>in: Law and Society Review,</journal>
+        <volume>vol. 9,</volume>
+        <pages>pp. 95—160.</pages>
+    </sequence>
+    <sequence>
+        <author>Geiger, Theodor,</author>
+        <date>1964:</date>
+        <title>Vorstudien zu einer Soziologie des Rechts.</title>
+        <author>Neuwied. Gessner, Volkmar,</author>
+        <date>1976:</date>
+        <title>Recht und Konflikt.</title>
+        <location>Tübingen.</location>
+    </sequence>
+    <sequence>
+        <location>Hilden, Hartmut,</location>
+        <date>1976:</date>
+        <title>Rechtstatsachen im Räumungsstreit. Frankfurt/Main.</title>
+    </sequence>
+    <sequence>
+        <author>Johnson, Earl,</author>
+        <date>1979;</date>
+        <title>Thinking about Access: A Preliminary Typology of Possible Strategies.</title>
+        <editor>In: Cappelletti, Mauro und Bryant Garth (Hrsg.):</editor>
+        <container-title>Access to Justice,</container-title>
+        <volume>Bd. III.</volume>
+        <editor>Milan und Alphen/ Rijn.</editor>
+    </sequence>
+    <sequence>
+        <author>Koch, Hartmut,</author>
+        <date>1975:</date>
+        <title>Das Gerichtsverfahren als Konfliktlösungsprozeß — Einstellung von Klägern und Beklagten zu Mietprozessen.</title>
+        <location>Diss. Hamburg.</location>
+    </sequence>
+    <sequence>
+        <author>Luhmann, Niklas,</author>
+        <date>1969:</date>
+        <title>Legitimation durch Verfahren.</title>
+        <location>Neuwied und Darmstadt.</location>
+    </sequence>
+    <sequence>
+        <author>Luhmann, Niklas,</author>
+        <date>1980:</date>
+        <title>Kommunikation über Recht in Interaktionssystemen,</title>
+        <editor>in: Blankenburg, Erhard; Klausa, Ekkehard und Hubert Rottleuthner (Hrsg.):</editor>
+        <container-title>Alternative Rechtsformen und Alternativen zum Recht,</container-title>
+        <collection-title>Jahrbuch für Rechtssoziologie und Rechtstheorie</collection-title>
+        <volume>Bd. 6. Opladen.</volume>
+    </sequence>
+    <sequence>
+        <author>Reifner, Udo,</author>
+        <date>1978:</date>
+        <title>Rechtshilfebedürfnis und Verrechtlichung am Beispiel einer Berliner Mieterinitiative,</title>
+        <location>Wissenschaftszentrum Berlin,</location>
+        <pages>IIM-dp/78—70.</pages>
+    </sequence>
+    <sequence>
+        <author>Reifner, Udo,</author>
+        <date>1979:</date>
+        <title>Gewerkschaftlicher Rechtsschutz — Geschichte des freigewerkschaftlichen Rechtsschutzes und der Rechtsberatung der Deutschen Arbeitsfront von 1894—1945.</title>
+        <location>Wissenschaftszentrum Berlin</location>
+        <pages>IIM-dp/79—104.</pages>
+    </sequence>
+    <sequence>
+        <author>Reifner,</author>
+        <title>Udo und Irmela Gorges,</title>
+        <date>1980;</date>
+        <title>Alternativen der Rechtsberatung: Dienstleistung, Fürsorge und kollektive Selbsthilfe,</title>
+        <editor>in: Blankenburg, Erhard; Klausa, Ekkehard und Hubert Rottleuthner (Hrsg.):</editor>
+        <container-title>Alternative Rechtsformen und Alternativen zum Recht,</container-title>
+        <collection-title>Jahrbuch für Rechtssoziologie und Rechtstheorie</collection-title>
+        <volume>Bd. 6. Opladen.</volume>
+    </sequence>
+    <sequence>
+        <author>Sarat,</author>
+        <location>Austin,</location>
+        <date>1976:</date>
+        <title>Alternatives in Dispute Processing in a Small Claim Court,</title>
+        <journal>in: Law and Society Review,</journal>
+        <volume>vol. 10,</volume>
+        <pages>pp. 339—375.</pages>
+    </sequence>
+    <sequence>
+        <author>Schönholz, Siegfried,</author>
+        <date>1980:</date>
+        <title>Arbeitsplatzsicherung oder Kompensation - Rechtliche Formen des Bestandsschutzes im Vergleich</title>
+        <editor>in: Vereinigung für Rechtssoziologie (Hrsg.):</editor>
+        <container-title>Arbeitslosigkeit und Recht.</container-title>
+        <location>Baden-Baden.</location>
+    </sequence>
+    <sequence>
+        <author>Steinbach, E.,</author>
+        <date>1979:</date>
+        <title>GMD-Bericht, Manuskript. Gesellschaft für Mathematik und Datenverarbeitung.</title>
+        <location>Birlinghoven bei Bonn.</location>
+    </sequence>
+    <sequence>
+        <author>Wanner, Craig,</author>
+        <date>1975:</date>
+        <title>The Public Ordering of Private Cases; Winning Civil Court Cases,</title>
+        <journal>in: Law and Society Review,</journal>
+        <volume>vol. 9,</volume>
+        <pages>pp. 293-306.</pages>
+    </sequence>
+</dataset>
--- a/prodigy/in/10.1515_zfrs-1980-0104.ttx
+++ b/prodigy/in/10.1515_zfrs-1980-0104.ttx
--- a/prodigy/in/10.1515_zfrs-1980-0104.xml
+++ b/prodigy/in/10.1515_zfrs-1980-0104.xml
--- a/prodigy/out/10.1515_zfrs-1980-0103-finder.jsonl
+++ b/prodigy/out/10.1515_zfrs-1980-0103-finder.jsonl
--- a/prodigy/out/10.1515_zfrs-1980-0103-parser.jsonl
+++ b/prodigy/out/10.1515_zfrs-1980-0103-parser.jsonl
--- a/prodigy/out/10.1515_zfrs-1980-0104-finder.jsonl
+++ b/prodigy/out/10.1515_zfrs-1980-0104-finder.jsonl
--- a/prodigy/out/10.1515_zfrs-1980-0104-parser.jsonl
+++ b/prodigy/out/10.1515_zfrs-1980-0104-parser.jsonl