git commit -m "Promotion: fix(tag-matching) | Fehlerhafte Zählung durch unvollständiges Matching von zusammengesetzten Tags behoben (Nummer:Typ:Begriff), semantische Inkonsistenzen (z. B. 'digitales:lernen' statt 'digital:lernen') korrigiert"

analyse_korrelation.py

@@ -1,3 +1,4 @@
+
 import os
 
 # Neue Exportfunktion: HTML in /tmp speichern, per SCP übertragen, PNG lokal speichern
@@ -288,7 +289,6 @@ def visualize_bivariate_correlation(df, x_terms, y_terms, title, x_label, y_labe
                             abs_corr = abs(corr)
                             significance = 'Signifikant' if p_value < 0.05 else 'Nicht signifikant'
                             hover_color = colors['brightArea'] if p_value < 0.05 else colors['depthArea']
-                            cooccurrence_count = int(((df[x_term] == 1) & (df[y_term] == 1)).sum())
                             correlations.append({
                                 'x_term': x_term,
                                 'y_term': y_term,
@@ -297,8 +297,6 @@ def visualize_bivariate_correlation(df, x_terms, y_terms, title, x_label, y_labe
                                 'p_value': p_value,
                                 'significance': significance,
                                 'hover_color': hover_color,
-                                'n_observations': int(len(x_valid)),
-                                'cooccurrence_count': cooccurrence_count,
                                 'interpretation': (
                                     f"Die Korrelation zwischen '{x_term}' und '{y_term}' beträgt {corr:.2f}. "
                                     f"p-Wert: {p_value:.3e} ({significance})"
@@ -364,25 +362,12 @@ def visualize_bivariate_correlation(df, x_terms, y_terms, title, x_label, y_labe
             line=dict(width=1, color=colors['background'])
         ),
         hovertemplate=(
-            '<b>%{customdata[0]}</b> ↔ <b>%{customdata[1]}</b><br>'
+            '<b>%{customdata[0]}</b><br>'
             'Korrelation: %{marker.color:.2f}<br>'
-            'p-Wert: %{customdata[3]:.3e}<br>'
+            'p-Wert: %{customdata[1]:.3e}<br>'
-            'Signifikanz: %{customdata[4]}<br>'
+            'Signifikanz: %{customdata[2]}'
-            'Stichprobe (n): %{customdata[5]}<br>'
-            'Gemeinsame Treffer: %{customdata[6]}<br>'
-            '%{customdata[7]}'
-            '<extra></extra>'
         ),
-        customdata=np.array(list(zip(
+        customdata=correlation_df[['x_term', 'p_value', 'significance']].to_numpy()
-            correlation_df['x_term'],
-            correlation_df['y_term'],
-            correlation_df['correlation'],
-            correlation_df['p_value'],
-            correlation_df['significance'],
-            correlation_df['n_observations'],
-            correlation_df['cooccurrence_count'],
-            correlation_df['interpretation']
-        )), dtype=object)
     )
 
     # Standardlayout verwenden und ggf. ergänzen, Margin dynamisch für Responsivität
@@ -392,7 +377,6 @@ def visualize_bivariate_correlation(df, x_terms, y_terms, title, x_label, y_labe
             x_title=x_label,
             y_title=y_label
         ),
-        hovermode='closest',
         xaxis=dict(
             tickangle=-45,
             automargin=True
@@ -507,6 +491,11 @@ df['X_Dimension'] = df[[tag for tag in tags_to_search_processed if tag in df.col
 df['Y_Dimension'] = df[[cat for cat in categories_processed if cat in df.columns]].sum(axis=1)
 df['Z_Dimension'] = df[[rq for rq in research_questions_processed if rq in df.columns]].sum(axis=1)
 
+# Clusteranalyse mit K-Means basierend auf den deduktiven Dimensionen
+features = df[['X_Dimension', 'Y_Dimension', 'Z_Dimension']]
+scaler = StandardScaler()
+scaled_features = scaler.fit_transform(features)
+
 # Clusteranalyse mit K-Means basierend auf den deduktiven Dimensionen
 # Prüfung auf konstante deduktive Dimensionen
 if df[['X_Dimension', 'Y_Dimension', 'Z_Dimension']].nunique().eq(1).all():
@@ -579,19 +568,6 @@ for cluster in cluster_means.index:
 
 # Statische Cluster-Beschriftungen in den DataFrame einfügen
 df['Cluster_Label'] = df['KMeans_Cluster'].map(cluster_labels)
-df['Cluster_Label'] = df['Cluster_Label'].fillna(df['KMeans_Cluster'])
-
-# Farbzuordnung für die Clusterlabels aus den CI-Farben ableiten
-fallback_color = cluster_colors.get("0", colors.get('primaryLine', '#1f77b4'))
-color_map = {}
-for cluster_key, label in cluster_labels.items():
-    base_color = cluster_colors.get(str(cluster_key), fallback_color)
-    color_map[label] = base_color
-
-# Sicherstellen, dass auch eventuelle Restlabels (z.B. "Nicht gültig") erfasst werden
-for label in df['Cluster_Label'].dropna().unique():
-    if label not in color_map:
-        color_map[label] = cluster_colors.get(str(label), fallback_color)
 
 # Ausgabe der statischen Cluster-Beschriftungen
 print("Cluster-Beschriftungen (inhaltlich):")
@@ -608,7 +584,7 @@ fig_cluster = px.scatter_3d(
     color='Cluster_Label',
     size='Point_Size',
     size_max=100,
-    color_discrete_map=color_map,
+    color_discrete_sequence=list(cluster_colors.values()),
     hover_data={
         'Cluster_Label': True,
         'X_Dimension': True,
@@ -626,6 +602,7 @@ fig_cluster = px.scatter_3d(
     }
 )
 
+
 # Layout mit Standardlayout und konsistenten CI-konformen Ergänzungen
 layout_cluster = get_standard_layout(
     title=plot_title,
@@ -710,7 +687,6 @@ correlation_quality_results = {
     "Forschungsunterfragen & Kategorien": analyze_correlation_quality(df, research_questions_processed, categories_processed),
     "Forschungsunterfragen & Suchbegriffe": analyze_correlation_quality(df, research_questions_processed, tags_to_search_processed),
     "Forschungsunterfragen & Indizes": analyze_correlation_quality(df, research_questions_processed, index_terms_processed),
-    "Forschungsunterfragen & Forschungsunterfragen": analyze_correlation_quality(df, research_questions_processed, research_questions_processed),
     "Indizes & Kategorien": analyze_correlation_quality(df, index_terms_processed, categories_processed),
     "Indizes & Suchbegriffe": analyze_correlation_quality(df, index_terms_processed, tags_to_search_processed),
     "Suchbegriffe & Kategorien": analyze_correlation_quality(df, tags_to_search_processed, categories_processed),
@@ -777,17 +753,8 @@ def plot_average_correlation_plotly(summary_df):
     )
     # PNG-Export ergänzen
     png_path = os.path.join(export_path_png, f"{slugify('summary_plot_' + global_bib_filename.replace('.bib', ''))}.png")
-    try:
     fig.write_image(png_path, width=1200, height=800, scale=2)
     print(f"✅ PNG-Summary-Datei gespeichert unter: {png_path}")
-    except ValueError as err:
-        if "kaleido" in str(err).lower():
-            print("⚠️ PNG-Export übersprungen: Plotly benötigt das Paket 'kaleido'.")
-            print("   Installation (falls gewünscht): pip install -U kaleido")
-        else:
-            print(f"⚠️ PNG-Export fehlgeschlagen: {err}")
-    except Exception as err:
-        print(f"⚠️ PNG-Export fehlgeschlagen: {err}")
 
 #============================
 # Aufruf Alle möglichen bivariaten Korrelationen visualisieren

analyse_netzwerk.py

@@ -1,6 +1,7 @@
 
 from config_netzwerk import theme, export_fig_visual, bib_filename
 
+
 import os
 
 # Clear the terminal
@@ -59,14 +60,11 @@ from config_netzwerk import (
     export_fig_visualize_sources_status,
     export_fig_create_wordcloud_from_titles,
     export_fig_visualize_languages,
-    export_fig_visualize_relevance_fu,
-    export_fig_visualize_relevance_categories,
-    export_fig_visualize_relevance_search_terms,
 )
 
 from config_netzwerk import export_fig_png
 
-def export_figure_local(fig, name, flag):
+def export_figure_local(fig, name, flag, bib_filename=None):
     from config_netzwerk import export_path_html, export_path_png
     # Einmalige Definition von safe_filename am Anfang der Funktion
     safe_filename = prepare_figure_export(fig, name).replace(".html", "")
@@ -102,23 +100,6 @@ word_colors = [
     colors["negativeHighlight"]
 ]
 
-# Relevanz-Stufen (1 = gering, 5 = sehr hoch)
-RELEVANCE_LEVELS = [5, 4, 3, 2, 1]
-RELEVANCE_LEVEL_LABELS = {
-    5: "Relevanz 5",
-    4: "Relevanz 4",
-    3: "Relevanz 3",
-    2: "Relevanz 2",
-    1: "Relevanz 1",
-}
-RELEVANCE_COLOR_MAP = {
-    "Relevanz 5": colors['positiveHighlight'],
-    "Relevanz 4": colors['accent'],
-    "Relevanz 3": colors['brightArea'],
-    "Relevanz 2": colors['depthArea'],
-    "Relevanz 1": colors['negativeHighlight'],
-}
-
 # Aktuelles Datum
 current_date = datetime.now().strftime("%Y-%m-%d")
 
@@ -139,13 +120,6 @@ with open('en_complete.txt', 'r', encoding='utf-8') as file:
 # Kombinierte Stoppliste
 stop_words = stop_words_de.union(stop_words_en)
 
-# Hilfsfunktion: Relevanzstufe aus Keywords extrahieren
-def extract_relevance_level(entry_keywords):
-    for level in RELEVANCE_LEVELS:
-        if f'promotion:relevanz:{level}' in entry_keywords:
-            return level
-    return None
-
 # Funktion zur Berechnung der Stichprobengröße
 def calculate_sample_size(N, Z=1.96, p=0.5, e=0.05):
     n_0 = (Z**2 * p * (1 - p)) / (e**2)
@@ -175,10 +149,8 @@ def visualize_network(bib_database):
         'Buch',
         'Buchteil',
         'Bericht',
-        'Konferenz-Paper',
+        'Konferenz-Paper'
-        'Studienbrief'
     ]
-    
     tags_to_search = set()
     for number, type_ in product(numbers, types):
         search_term = search_terms[number]
@@ -314,15 +286,11 @@ def visualize_network(bib_database):
     secondary_nodes = []
     tertiary_nodes = []
 
-    total_fundzahlen = sum(fundzahlen.values())
-
     for node in G.nodes():
         color = G.nodes[node]['color']
         size = math.log(G.nodes[node].get('size', 10) + 1) * 10
         x, y = pos[node]
-        count = fundzahlen.get(node, 0)
+        hovertext = f"{node}<br>Anzahl Funde: {fundzahlen.get(node, 0)}"
-        percentage = (count / total_fundzahlen * 100) if total_fundzahlen else 0
-        hovertext = f"{node}<br>Anzahl Funde: {count}<br>Anteil: {percentage:.1f}%"
         node_data = dict(x=x, y=y, text=node, size=size, hovertext=hovertext)
         if color == colors['primaryLine']:
             primary_nodes.append(node_data)
@@ -363,7 +331,7 @@ def visualize_network(bib_database):
 
     fig = go.Figure(data=[edge_trace, primary_trace, secondary_trace, tertiary_trace])
     layout = get_standard_layout(
-        title=f"Suchbegriff-Netzwerk nach Relevanz und Semantik (n={total_fundzahlen}, Stand: {current_date})",
+        title=f"Suchbegriff-Netzwerk nach Relevanz und Semantik (n={sum(fundzahlen.values())}, Stand: {current_date})",
         x_title="Technologische Dimension",
         y_title="Pädagogische Dimension"
     )
@@ -375,7 +343,7 @@ def visualize_network(bib_database):
     fig.update_layout(**layout)
 
     fig.show(config={"responsive": True})
-    export_figure_local(fig, "visualize_network", export_fig_visualize_network)
+    export_figure_local(fig, "visualize_network", export_fig_visualize_network, bib_filename)
 
     # Einfache Pfadanalyse nach dem Anzeigen der Figur
     if 'e-learning' in G and 'online:lernen' in G:
@@ -410,11 +378,10 @@ def visualize_tags(bib_database):
         'Buch',
         'Buchteil',
         'Bericht',
-        'Konferenz-Paper',
+        'Konferenz-Paper'
-        'Studienbrief'
     ]
     tags_to_search = set(
-        f"#{number}:{type_}:{search_terms[number]}".lower()
+        f"#{number}:{type_}:{search_terms[number]}"
         for number, type_ in product(numbers, types)
     )
 
@@ -436,46 +403,36 @@ def visualize_tags(bib_database):
                         tag_counts[tag] += 1
 
     # Daten für Visualisierung aufbereiten
-    data_rows = [
+    data = [
-        {
+        {'Tag': tag, 'Count': count, 'Type': tag.split(':')[1].lower()}
-            'Tag': tag,
-            'Count': count,
-            'Type': tag.split(':')[1].lower()
-        }
         for tag, count in tag_counts.items()
         if count > 0
     ]
 
-    if not data_rows:
+    if not data:
         print("Warnung: Keine Tags gefunden, die den Suchkriterien entsprechen.")
         return
 
-    df = pd.DataFrame(data_rows)
-    df['TypeLabel'] = df['Type'].str.replace('-', ' ').str.title()
-    total_count = df['Count'].sum()
-    df['Percentage'] = df['Count'] / total_count * 100 if total_count else 0
-
     # Farbzuordnung
     color_map = {
         'zeitschriftenartikel': colors['primaryLine'],
         'konferenz-paper': colors['secondaryLine'],
         'buch': colors['depthArea'],
         'buchteil': colors['brightArea'],
-        'bericht': colors['accent'],
+        'bericht': colors['accent']
-        'studienbrief': colors['positiveHighlight']
     }
 
     # Visualisierung erstellen
+    total_count = sum(tag_counts.values())
     fig = px.bar(
-        df,
+        data,
         x='Tag',
         y='Count',
         title=f'Häufigkeit der Suchbegriffe in der Literaturanalyse (n={total_count}, Stand: {current_date})',
         labels={'Tag': 'Tag', 'Count': 'Anzahl der Vorkommen'},
         color='Type',
         color_discrete_map=color_map,
-        text_auto=True,
+        text_auto=True
-        custom_data=['TypeLabel', 'Percentage']
     )
 
     layout = get_standard_layout(
@@ -491,17 +448,9 @@ def visualize_tags(bib_database):
     layout["xaxis"]["automargin"] = True
     layout["autosize"] = True
     fig.update_layout(**layout)
-    fig.update_traces(
-        hovertemplate=(
-            "<b>%{x}</b><br>"
-            "Typ: %{customdata[0]}<br>"
-            "Anzahl: %{y}<br>"
-            "Anteil: %{customdata[1]:.1f}%<extra></extra>"
-        )
-    )
 
     fig.show(config={"responsive": True})
-    export_figure_local(fig, "visualize_tags", export_fig_visualize_tags)
+    export_figure_local(fig, "visualize_tags", export_fig_visualize_tags, bib_filename)
 
  # Visualisierung 3: Häufigkeit Index
 def visualize_index(bib_database):
@@ -529,21 +478,11 @@ def visualize_index(bib_database):
     index_data = [{'Index': index, 'Count': count} for index, count in index_counts.items()]
     index_data = sorted(index_data, key=lambda x: x['Count'], reverse=True)
 
-    index_df = pd.DataFrame(index_data)
+    total_count = sum(index_counts.values())
-    total_count = index_df['Count'].sum()
-    index_df['Percentage'] = index_df['Count'] / total_count * 100 if total_count else 0
     print(f"Häufigkeit Indizes (Gesamtanzahl: {total_count}):")
-    print(tabulate(index_df.to_dict('records'), headers="keys", tablefmt="grid"))
+    print(tabulate(index_data, headers="keys", tablefmt="grid"))
 
-    fig = px.bar(
+    fig = px.bar(index_data, x='Index', y='Count', title=f'Relevanzschlüssel nach Indexkategorien (n={total_count}, Stand: {current_date})', labels={'Index': 'Index', 'Count': 'Anzahl der Vorkommen'}, text_auto=True)
-        index_df,
-        x='Index',
-        y='Count',
-        title=f'Relevanzschlüssel nach Indexkategorien (n={total_count}, Stand: {current_date})',
-        labels={'Index': 'Index', 'Count': 'Anzahl der Vorkommen'},
-        text_auto=True,
-        custom_data=['Percentage']
-    )
     layout = get_standard_layout(
         title=fig.layout.title.text,
         x_title='Index',
@@ -558,15 +497,8 @@ def visualize_index(bib_database):
     layout["autosize"] = True
     fig.update_layout(**layout)
     fig.update_traces(marker=plot_styles['balken_primaryLine'])
-    fig.update_traces(
-        hovertemplate=(
-            "<b>%{x}</b><br>"
-            "Anzahl: %{y}<br>"
-            "Anteil: %{customdata[0]:.1f}%<extra></extra>"
-        )
-    )
     fig.show(config={"responsive": True})
-    export_figure_local(fig, "visualize_index", export_fig_visualize_index)
+    export_figure_local(fig, "visualize_index", export_fig_visualize_index, bib_filename)
 
  # Visualisierung 4: Häufigkeit Forschungsunterfragen
 def visualize_research_questions(bib_database):
@@ -593,22 +525,13 @@ def visualize_research_questions(bib_database):
     rq_data = [{'Research_Question': research_questions[keyword], 'Count': count} for keyword, count in rq_counts.items()]
     rq_data = sorted(rq_data, key=lambda x: x['Count'], reverse=True)
 
-    rq_data_df = pd.DataFrame(rq_data, columns=['Research_Question', 'Count'])
+    rq_data_df = pd.DataFrame(rq_data)
 
     total_count = rq_data_df['Count'].sum()
-    rq_data_df['Percentage'] = rq_data_df['Count'] / total_count * 100 if total_count else 0
     print(f"Häufigkeit Forschungsunterfragen (Gesamtanzahl: {total_count}):")
     print(tabulate(rq_data, headers="keys", tablefmt="grid"))
 
-    fig = px.bar(
+    fig = px.bar(rq_data_df, x='Research_Question', y='Count', title=f'Zuordnung der Literatur zu Forschungsunterfragen (n={total_count}, Stand: {current_date})', labels={'Research_Question': 'Forschungsunterfrage', 'Count': 'Anzahl der Vorkommen'}, text_auto=True)
-        rq_data_df,
-        x='Research_Question',
-        y='Count',
-        title=f'Zuordnung der Literatur zu Forschungsunterfragen (n={total_count}, Stand: {current_date})',
-        labels={'Research_Question': 'Forschungsunterfrage', 'Count': 'Anzahl der Vorkommen'},
-        text_auto=True,
-        custom_data=['Percentage']
-    )
     layout = get_standard_layout(
         title=fig.layout.title.text,
         x_title='Forschungsunterfrage',
@@ -623,15 +546,8 @@ def visualize_research_questions(bib_database):
     layout["autosize"] = True
     fig.update_layout(**layout)
     fig.update_traces(marker=plot_styles['balken_primaryLine'])
-    fig.update_traces(
-        hovertemplate=(
-            "<b>%{x}</b><br>"
-            "Anzahl: %{y}<br>"
-            "Anteil: %{customdata[0]:.1f}%<extra></extra>"
-        )
-    )
     fig.show(config={"responsive": True})
-    export_figure_local(fig, "visualize_research_questions", export_fig_visualize_research_questions)
+    export_figure_local(fig, "visualize_research_questions", export_fig_visualize_research_questions, bib_filename)
 
  # Visualisierung 5: Häufigkeit spezifischer Kategorien
 def visualize_categories(bib_database):
@@ -653,22 +569,13 @@ def visualize_categories(bib_database):
     cat_data = [{'Category': categories[keyword], 'Count': count} for keyword, count in cat_counts.items()]
     cat_data = sorted(cat_data, key=lambda x: x['Count'], reverse=True)
 
-    cat_data_df = pd.DataFrame(cat_data, columns=['Category', 'Count'])
+    cat_data_df = pd.DataFrame(cat_data)
 
     total_count = cat_data_df['Count'].sum()
-    cat_data_df['Percentage'] = cat_data_df['Count'] / total_count * 100 if total_count else 0
     print(f"Häufigkeit Kategorien (Gesamtanzahl: {total_count}):")
     print(tabulate(cat_data, headers="keys", tablefmt="grid"))
 
-    fig = px.bar(
+    fig = px.bar(cat_data_df, x='Category', y='Count', title=f'Textsortenzuordnung der analysierten Quellen (n={total_count}, Stand: {current_date})', labels={'Category': 'Kategorie', 'Count': 'Anzahl der Vorkommen'}, text_auto=True)
-        cat_data_df,
-        x='Category',
-        y='Count',
-        title=f'Textsortenzuordnung der analysierten Quellen (n={total_count}, Stand: {current_date})',
-        labels={'Category': 'Kategorie', 'Count': 'Anzahl der Vorkommen'},
-        text_auto=True,
-        custom_data=['Percentage']
-    )
     layout = get_standard_layout(
         title=fig.layout.title.text,
         x_title='Kategorie',
@@ -683,179 +590,8 @@ def visualize_categories(bib_database):
     layout["autosize"] = True
     fig.update_layout(**layout)
     fig.update_traces(marker=plot_styles['balken_primaryLine'])
-    fig.update_traces(
-        hovertemplate=(
-            "<b>%{x}</b><br>"
-            "Anzahl: %{y}<br>"
-            "Anteil: %{customdata[0]:.1f}%<extra></extra>"
-        )
-    )
     fig.show(config={"responsive": True})
-    export_figure_local(fig, "visualize_categories", export_fig_visualize_categories)
+    export_figure_local(fig, "visualize_categories", export_fig_visualize_categories, bib_filename)
-
-# Relevanz-Auswertungen
-def build_relevance_distribution(bib_database, tag_to_label):
-    records = []
-
-    for entry in bib_database.entries:
-        keywords_raw = entry.get('keywords', '')
-        if not keywords_raw:
-            continue
-
-        entry_keywords = set(map(str.lower, map(str.strip, keywords_raw.replace('\\#', '#').split(','))))
-        relevance_level = extract_relevance_level(entry_keywords)
-        if relevance_level is None:
-            continue
-
-        for tag, label in tag_to_label.items():
-            if tag in entry_keywords:
-                records.append({
-                    'Kategorie': label,
-                    'Relevanzstufe': RELEVANCE_LEVEL_LABELS[relevance_level]
-                })
-
-    if not records:
-        return pd.DataFrame()
-
-    df = pd.DataFrame(records)
-    df = (
-        df.groupby(['Kategorie', 'Relevanzstufe'])
-        .size()
-        .reset_index(name='Count')
-    )
-    df['Relevanzstufe'] = pd.Categorical(
-        df['Relevanzstufe'],
-        categories=[RELEVANCE_LEVEL_LABELS[level] for level in RELEVANCE_LEVELS],
-        ordered=True
-    )
-    return df.sort_values(['Kategorie', 'Relevanzstufe'])
-
-
-def plot_relevance_distribution(df, title, x_title, export_flag, filename):
-    if df.empty:
-        print(f"⚠️ Keine Relevanzdaten verfügbar für: {title}")
-        return
-
-    total_count = df['Count'].sum()
-    df['Percentage'] = df['Count'] / total_count * 100 if total_count else 0
-    fig = px.bar(
-        df,
-        x='Kategorie',
-        y='Count',
-        color='Relevanzstufe',
-        color_discrete_map=RELEVANCE_COLOR_MAP,
-        category_orders={'Relevanzstufe': [RELEVANCE_LEVEL_LABELS[level] for level in RELEVANCE_LEVELS]},
-        title=f"{title} (n={total_count}, Stand: {current_date})",
-        labels={'Kategorie': x_title, 'Count': 'Anzahl', 'Relevanzstufe': 'Relevanzstufe'},
-        custom_data=['Relevanzstufe', 'Percentage']
-    )
-
-    layout = get_standard_layout(
-        title=fig.layout.title.text,
-        x_title=x_title,
-        y_title='Anzahl'
-    )
-    layout['barmode'] = 'stack'
-    layout['font'] = {"size": 14, "color": colors['text']}
-    layout['title'] = {"font": {"size": 16}}
-    layout['margin'] = dict(b=160, t=60, l=40, r=40)
-    layout['xaxis'] = layout.get('xaxis', {})
-    layout['xaxis']['tickangle'] = -45
-    layout['xaxis']['automargin'] = True
-    layout['autosize'] = True
-    fig.update_layout(**layout)
-    fig.update_traces(
-        hovertemplate=(
-            "<b>%{x}</b><br>"
-            "Relevanzstufe: %{customdata[0]}<br>"
-            "Anzahl: %{y}<br>"
-            "Anteil: %{customdata[1]:.1f}%<extra></extra>"
-        )
-    )
-
-    fig.show(config={"responsive": True})
-    export_figure_local(fig, filename, export_flag)
-
-
-def visualize_relevance_vs_research_questions(bib_database):
-    research_questions = {
-        'promotion:fu1': 'Akzeptanz und Nützlichkeit (FU1)',
-        'promotion:fu2a': 'Effekt für Lernende (FU2a)',
-        'promotion:fu2b': 'Effekt-Faktoren für Lehrende (FU2b)',
-        'promotion:fu3': 'Konzeption und Merkmale (FU3)',
-        'promotion:fu4a': 'Bildungswissenschaftliche Mechanismen (FU4a)',
-        'promotion:fu4b': 'Technisch-gestalterische Mechanismen (FU4b)',
-        'promotion:fu5': 'Möglichkeiten und Grenzen (FU5)',
-        'promotion:fu6': 'Beurteilung als Kompetenzerwerbssystem (FU6)',
-        'promotion:fu7': 'Inputs und Strategien (FU7)'
-    }
-    tag_to_label = {key.lower(): value for key, value in research_questions.items()}
-    df = build_relevance_distribution(bib_database, tag_to_label)
-    plot_relevance_distribution(
-        df,
-        "Relevanzverteilung nach Forschungsunterfragen",
-        "Forschungsunterfragen",
-        export_fig_visualize_relevance_fu,
-        "visualize_relevance_fu"
-    )
-
-
-def visualize_relevance_vs_categories(bib_database):
-    categories = {
-        'promotion:argumentation': 'Argumentation',
-        'promotion:kerngedanke': 'Kerngedanke',
-        'promotion:weiterführung': 'Weiterführung',
-        'promotion:schlussfolgerung': 'Schlussfolgerung'
-    }
-    tag_to_label = {key.lower(): value for key, value in categories.items()}
-    df = build_relevance_distribution(bib_database, tag_to_label)
-    plot_relevance_distribution(
-        df,
-        "Relevanzverteilung nach Kategorien",
-        "Kategorien",
-        export_fig_visualize_relevance_categories,
-        "visualize_relevance_categories"
-    )
-
-
-def visualize_relevance_vs_search_terms(bib_database):
-    search_terms = {
-        '0': 'digital:learning',
-        '1': 'learning:management:system',
-        '2': 'online:lernplattform',
-        '3': 'online:lernumgebung',
-        '4': 'mooc',
-        '5': 'e-learning',
-        '6': 'bildung:technologie',
-        '7': 'digital:medien',
-        '8': 'blended:learning',
-        '9': 'digital:lernen',
-        'a': 'online:lernen',
-        'b': 'online:learning'
-    }
-    types = [
-        'Zeitschriftenartikel',
-        'Buch',
-        'Buchteil',
-        'Bericht',
-        'Konferenz-Paper',
-        'Studienbrief'
-    ]
-
-    tag_to_label = {}
-    for number, term in search_terms.items():
-        for type_ in types:
-            tag = f'#{number}:{type_}:{term}'.lower()
-            tag_to_label[tag] = f"#{number}:{term}"
-
-    df = build_relevance_distribution(bib_database, tag_to_label)
-    plot_relevance_distribution(
-        df,
-        "Relevanzverteilung nach Suchbegriffen",
-        "Suchbegriffe",
-        export_fig_visualize_relevance_search_terms,
-        "visualize_relevance_search_terms"
-    )
 
  # Zeitreihenanalyse der Veröffentlichungen
 def visualize_time_series(bib_database):
@@ -878,16 +614,13 @@ def visualize_time_series(bib_database):
     if publication_years:
         year_counts = Counter(publication_years)
         df = pd.DataFrame(year_counts.items(), columns=['Year', 'Count']).sort_values('Year')
-        total_publications = df['Count'].sum()
-        df['Percentage'] = df['Count'] / total_publications * 100 if total_publications else 0
 
         fig = px.line(
             df,
             x='Year',
             y='Count',
             title=f'Jährliche Veröffentlichungen in der Literaturanalyse (n={sum(year_counts.values())}, Stand: {current_date})',
-            labels={'Year': 'Jahr', 'Count': 'Anzahl der Veröffentlichungen'},
+            labels={'Year': 'Jahr', 'Count': 'Anzahl der Veröffentlichungen'}
-            custom_data=['Percentage']
         )
         layout = get_standard_layout(
             title=fig.layout.title.text,
@@ -904,15 +637,8 @@ def visualize_time_series(bib_database):
         layout["autosize"] = True
         fig.update_layout(**layout)
         fig.update_traces(line=plot_styles['linie_primaryLine'])
-        fig.update_traces(
-            hovertemplate=(
-                "<b>%{x}</b><br>"
-                "Anzahl: %{y}<br>"
-                "Anteil: %{customdata[0]:.1f}%<extra></extra>"
-            )
-        )
         fig.show(config={"responsive": True})
-        export_figure_local(fig, "visualize_time_series", export_fig_visualize_time_series)
+        export_figure_local(fig, "visualize_time_series", export_fig_visualize_time_series, bib_filename)
     else:
         print("Keine gültigen Veröffentlichungsjahre gefunden.")
 
@@ -929,18 +655,8 @@ def visualize_top_authors(bib_database):
     top_authors = Counter(author_counts).most_common(top_n)
     if top_authors:
         df = pd.DataFrame(top_authors, columns=['Author', 'Count'])
-        overall_total = sum(author_counts.values())
-        df['Percentage'] = df['Count'] / overall_total * 100 if overall_total else 0
 
-        fig = px.bar(
+        fig = px.bar(df, x='Author', y='Count', title=f'Meistgenannte Autor:innen in der Literaturanalyse (Top {top_n}, n={sum(author_counts.values())}, Stand: {current_date})', labels={'Author': 'Autor', 'Count': 'Anzahl der Werke'}, text_auto=True)
-            df,
-            x='Author',
-            y='Count',
-            title=f'Meistgenannte Autor:innen in der Literaturanalyse (Top {top_n}, n={overall_total}, Stand: {current_date})',
-            labels={'Author': 'Autor', 'Count': 'Anzahl der Werke'},
-            text_auto=True,
-            custom_data=['Percentage']
-        )
         layout = get_standard_layout(
             title=fig.layout.title.text,
             x_title='Autor',
@@ -955,18 +671,55 @@ def visualize_top_authors(bib_database):
         layout["autosize"] = True
         fig.update_layout(**layout)
         fig.update_traces(marker=plot_styles['balken_primaryLine'])
-        fig.update_traces(
-            hovertemplate=(
-                "<b>%{x}</b><br>"
-                "Anzahl: %{y}<br>"
-                "Anteil: %{customdata[0]:.1f}%<extra></extra>"
-            )
-        )
         fig.show(config={"responsive": True})
-        export_figure_local(fig, "visualize_top_authors", export_fig_visualize_top_authors)
+        export_figure_local(fig, "visualize_top_authors", export_fig_visualize_top_authors, bib_filename)
     else:
         print("Keine Autoren gefunden.")
 
+ # Top Titel nach Anzahl der Werke
+def normalize_title(title):
+    # Entfernen von Sonderzeichen und Standardisierung auf Kleinbuchstaben
+    title = title.lower().translate(str.maketrans('', '', ",.!?\"'()[]{}:;"))
+    # Zusammenführen ähnlicher Titel, die sich nur in geringfügigen Details unterscheiden
+    title = " ".join(title.split())
+    # Entfernen häufiger Füllwörter oder Standardphrasen, die die Unterscheidung nicht unterstützen
+    common_phrases = ['eine studie', 'untersuchung der', 'analyse von']
+    for phrase in common_phrases:
+        title = title.replace(phrase, '')
+    return title.strip()
+
+def visualize_top_publications(bib_database):
+    top_n = 25  # Anzahl der Top-Publikationen, die angezeigt werden sollen
+    publication_counts = defaultdict(int)
+    
+    for entry in bib_database.entries:
+        if 'title' in entry:
+            title = normalize_title(entry['title'])
+            publication_counts[title] += 1
+
+    top_publications = sorted(publication_counts.items(), key=lambda x: x[1], reverse=True)[:top_n]
+    publication_data = [{'Title': title[:50] + '...' if len(title) > 50 else title, 'Count': count} for title, count in top_publications]
+
+    df = pd.DataFrame(publication_data)
+    
+    fig = px.bar(df, x='Title', y='Count', title=f'Häufig zitierte Publikationen in der Analyse (Top {top_n}, n={sum(publication_counts.values())}, Stand: {current_date})', labels={'Title': 'Titel', 'Count': 'Anzahl der Nennungen'})
+    layout = get_standard_layout(
+        title=fig.layout.title.text,
+        x_title='Titel',
+        y_title='Anzahl der Nennungen'
+    )
+    layout["font"] = {"size": 14, "color": colors['text']}
+    layout["title"] = {"font": {"size": 16}}
+    layout["margin"] = dict(b=160, t=60, l=40, r=40)
+    layout["xaxis"] = layout.get("xaxis", {})
+    layout["xaxis"]["tickangle"] = -45
+    layout["xaxis"]["automargin"] = True
+    layout["autosize"] = True
+    fig.update_layout(**layout)
+    fig.update_traces(marker=plot_styles['balken_primaryLine'])
+    fig.show(config={"responsive": True})
+    export_figure_local(fig, "visualize_top_publications", export_fig_visualize_top_publications, bib_filename)
+
 ##########
 
 # Daten vorbereiten
@@ -1008,8 +761,7 @@ def prepare_path_data(bib_database):
         'Buch',
         'Buchteil',
         'Bericht',
-        'Konferenz-Paper',
+        'Konferenz-Paper'
-        'Studienbrief'
     ]
 
     data = []
@@ -1048,14 +800,12 @@ def create_path_diagram(data):
     sources = []
     targets = []
     values = []
-    node_counts = Counter()
     color_map = {
         'zeitschriftenartikel': colors['primaryLine'],
         'konferenz-paper': colors['secondaryLine'],
         'buch': colors['depthArea'],
         'buchteil': colors['brightArea'],
-        'bericht': colors['accent'],
+        'bericht': colors['accent']
-        'studienbrief': colors['positiveHighlight']
     }
 
     def add_to_labels(label):
@@ -1072,19 +822,8 @@ def create_path_diagram(data):
         sources.extend([fu_idx, category_idx, index_idx])
         targets.extend([category_idx, index_idx, type_idx])
         values.extend([1, 1, 1])
-        node_counts.update([entry['FU'], entry['Category'], entry['Index'], entry['Type']])
 
     node_colors = [color_map.get(label, colors['primaryLine']) for label in labels]
-    total_paths = len(data)
-    total_flows = sum(values)
-    node_percentages = [
-        node_counts.get(label, 0) / total_paths * 100 if total_paths else 0
-        for label in labels
-    ]
-    link_percentages = [
-        value / total_flows * 100 if total_flows else 0
-        for value in values
-    ]
 
     fig = go.Figure(data=[go.Sankey(
         node=dict(
@@ -1092,24 +831,12 @@ def create_path_diagram(data):
             thickness=20,
             line=dict(color="black", width=0.5),
             label=labels,
-            color=node_colors,
+            color=node_colors
-            customdata=node_percentages,
-            hovertemplate=(
-                "%{label}<br>"
-                "Anzahl: %{value}<br>"
-                "Anteil der Pfade: %{customdata:.1f}%<extra></extra>"
-            )
         ),
         link=dict(
             source=sources,
             target=targets,
-            value=values,
+            value=values
-            customdata=link_percentages,
-            hovertemplate=(
-                "%{source.label} → %{target.label}<br>"
-                "Anzahl: %{value}<br>"
-                "Anteil der Verbindungen: %{customdata:.1f}%<extra></extra>"
-            )
         )
     )])
     layout = get_standard_layout(
@@ -1123,7 +850,7 @@ def create_path_diagram(data):
     layout["autosize"] = True
     fig.update_layout(**layout)
     fig.show(config={"responsive": True})
-    export_figure_local(fig, "create_path_diagram", export_fig_create_path_diagram)
+    export_figure_local(fig, "create_path_diagram", export_fig_create_path_diagram, bib_filename)
 
 #############
 
@@ -1227,54 +954,22 @@ def create_sankey_diagram(bib_database):
         colors['positiveHighlight']     # Ausgewählte Quellen
     ]
 
-    node_values = [
-        initial_sources,
-        screened_sources,
-        quality_sources,
-        relevance_sources,
-        thematic_sources,
-        recent_sources,
-        classic_sources,
-        selected_sources
-    ]
-    node_percentages = [
-        value / initial_sources * 100 if initial_sources else 0
-        for value in node_values
-    ]
-    link_percentages = [
-        value / initial_sources * 100 if initial_sources else 0
-        for value in values
-    ]
-
     # Sankey-Diagramm erstellen
     node_config = {
         **plot_styles["sankey_node"],
         "label": node_labels,
-        "color": node_colors,
+        "color": node_colors
-        "customdata": node_percentages,
-        "hovertemplate": (
-            "%{label}<br>"
-            "Anzahl: %{value}<br>"
-            "Anteil an Ausgangsmenge: %{customdata:.1f}%<extra></extra>"
-        )
     }
     # Remove any invalid 'font' key if present
     node_config.pop("font", None)
-    link_config = {
-        **plot_styles["sankey_link"],
-        "source": sources,
-        "target": targets,
-        "value": values,
-        "customdata": link_percentages,
-        "hovertemplate": (
-            "%{source.label} → %{target.label}<br>"
-            "Anzahl: %{value}<br>"
-            "Anteil an Ausgangsmenge: %{customdata:.1f}%<extra></extra>"
-        )
-    }
     fig = go.Figure(go.Sankey(
         node=node_config,
-        link=link_config
+        link=dict(
+            **plot_styles["sankey_link"],
+            source=sources,
+            target=targets,
+            value=values
+        )
     ))
     # Layout anpassen
     layout = get_standard_layout(
@@ -1288,7 +983,7 @@ def create_sankey_diagram(bib_database):
     layout["autosize"] = True
     fig.update_layout(**layout)
     fig.show(config={"responsive": True})
-    export_figure_local(fig, "create_sankey_diagram", export_fig_create_sankey_diagram)
+    export_figure_local(fig, "create_sankey_diagram", export_fig_create_sankey_diagram, bib_filename)
 
 ##########
 
@@ -1306,33 +1001,31 @@ def visualize_sources_status(bib_database):
     """
     Visualisiert den Status der analysierten und nicht analysierten Quellen pro Suchordner.
     """
-    search_terms = {
-        '0': 'digital:learning',
-        '1': 'learning:management:system',
-        '2': 'online:lernplattform',
-        '3': 'online:lernumgebung',
-        '4': 'mooc',
-        '5': 'e-learning',
-        '6': 'bildung:technologie',
-        '7': 'digital:medien',
-        '8': 'blended:learning',
-        '9': 'digital:lernen',
-        'a': 'online:lernen',
-        'b': 'online:learning'
-    }
-    numbers_order = ['0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 'a', 'b']
-    type_order = [
-        'Zeitschriftenartikel',
-        'Buch',
-        'Buchteil',
-        'Bericht',
-        'Konferenz-Paper',
-        'Studienbrief'
-    ]
     search_folder_tags = [
-        f"#{number}:{type_}:{search_terms[number]}".lower()
+        "#1:zeitschriftenartikel:learning:management:system",
-        for type_ in type_order
+        "#2:zeitschriftenartikel:online:lernplattform",
-        for number in numbers_order
+        "#3:zeitschriftenartikel:online:lernumgebung",
+        "#4:zeitschriftenartikel:mooc",
+        "#5:zeitschriftenartikel:e-learning",
+        "#6:zeitschriftenartikel:bildung:technologie",
+        "#7:zeitschriftenartikel:digital:medien",
+        "#8:zeitschriftenartikel:blended:learning",
+        "#9:zeitschriftenartikel:digital:lernen",
+        "#a:zeitschriftenartikel:online:lernen",
+        "#b:zeitschriftenartikel:online:learning",
+        "#0:zeitschriftenartikel:digital:learning",
+        "#1:konferenz-paper:learning:management:system",
+        "#2:konferenz-paper:online:lernplattform",
+        "#3:konferenz-paper:online:lernumgebung",
+        "#4:konferenz-paper:mooc",
+        "#5:konferenz-paper:e-learning",
+        "#6:konferenz-paper:bildung:technologie",
+        "#7:konferenz-paper:digital:medien",
+        "#8:konferenz-paper:blended:learning",
+        "#9:konferenz-paper:digital:lernen",
+        "#a:konferenz-paper:online:lernen",
+        "#b:konferenz-paper:online:learning",
+        "#0:konferenz-paper:digital:learning"
     ]
 
     category_tags = {"promotion:argumentation", "promotion:kerngedanke", "promotion:weiterführung", "promotion:schlussfolgerung"}
@@ -1387,45 +1080,21 @@ def visualize_sources_status(bib_database):
         tablefmt='grid'
     ))
 
-    total_identifiziert = sum(counts["Identifiziert"] for counts in source_data.values())
-    analysiert_percentages = [
-        value / total_identifiziert * 100 if total_identifiziert else 0
-        for value in analysiert_values
-    ]
-    nicht_analysiert_percentages = [
-        value / total_identifiziert * 100 if total_identifiziert else 0
-        for value in nicht_analysiert_values
-    ]
-
     fig = go.Figure()
     fig.add_trace(go.Bar(
         x=tags,
         y=analysiert_values,
         name='Analysiert',
-        marker=dict(color=analysiert_colors),
+        marker=dict(color=analysiert_colors)
-        customdata=analysiert_percentages,
-        hovertemplate=(
-            "<b>%{x}</b><br>"
-            "Status: Analysiert<br>"
-            "Anzahl: %{y}<br>"
-            "Anteil: %{customdata:.1f}%<extra></extra>"
-        )
     ))
     fig.add_trace(go.Bar(
         x=tags,
         y=nicht_analysiert_values,
         name='Nicht-Analysiert',
-        marker=plot_styles['balken_primaryLine'],
+        marker=plot_styles['balken_primaryLine']
-        customdata=nicht_analysiert_percentages,
-        hovertemplate=(
-            "<b>%{x}</b><br>"
-            "Status: Nicht-Analysiert<br>"
-            "Anzahl: %{y}<br>"
-            "Anteil: %{customdata:.1f}%<extra></extra>"
-        )
     ))
     layout = get_standard_layout(
-        title=f'Analyse- und Stichprobenstatus je Suchordner (n={total_identifiziert}, Stand: {current_date})',
+        title=f'Analyse- und Stichprobenstatus je Suchordner (n={sum(counts["Identifiziert"] for counts in source_data.values())}, Stand: {current_date})',
         x_title='Suchbegriffsordner',
         y_title='Anzahl der Quellen'
     )
@@ -1442,7 +1111,7 @@ def visualize_sources_status(bib_database):
     layout["autosize"] = True
     fig.update_layout(**layout)
     fig.show(config={"responsive": True})
-    export_figure_local(fig, "visualize_sources_status", export_fig_visualize_sources_status)
+    export_figure_local(fig, "visualize_sources_status", export_fig_visualize_sources_status, bib_filename)
 
 #############
 
@@ -1528,8 +1197,8 @@ def visualize_languages(bib_database):
         color='Gruppe',
         color_discrete_map=color_discrete_map,
         title=f'Sprachverteilung der analysierten Quellen (n={sum(norm_counts.values())}, Stand: {current_date})',
-        barmode="stack",
+        hover_data=["Sprache", "Gruppe", "Anzahl", "Anteil (%)"],
-        custom_data=['Gruppe', 'Anteil (%)']
+        barmode="stack"
     )
 
     layout = get_standard_layout(
@@ -1544,18 +1213,10 @@ def visualize_languages(bib_database):
     # Ergänzung: Y-Achse logarithmisch skalieren
     layout["yaxis_type"] = "log"
     fig.update_layout(**layout)
-    fig.update_traces(
-        hovertemplate=(
-            "<b>%{x}</b><br>"
-            "Sprachgruppe: %{customdata[0]}<br>"
-            "Anzahl: %{y}<br>"
-            "Anteil: %{customdata[1]:.2f}%<extra></extra>"
-        )
-    )
     fig.show(config={"responsive": True})
     # Tabelle ausgeben
     print(tabulate(df.sort_values("Anzahl", ascending=False), headers="keys", tablefmt="grid", showindex=False))
-    export_figure_local(fig, "visualize_languages", export_fig_visualize_languages)
+    export_figure_local(fig, "visualize_languages", export_fig_visualize_languages, bib_filename)
 
 # Visualisierung der Verteilung von ENTRYTYPE innerhalb jeder Sprache
 def visualize_language_entrytypes(bib_database):
@@ -1605,8 +1266,6 @@ def visualize_language_entrytypes(bib_database):
     grouped.rename(columns={'ENTRYTYPE': 'Eintragstyp'}, inplace=True)
     # Anteil innerhalb Sprache (%)
     grouped["Anteil innerhalb Sprache (%)"] = grouped.groupby("Sprache")["Anzahl"].transform(lambda x: (x / x.sum() * 100).round(2))
-    total_entrytypes = grouped['Anzahl'].sum()
-    grouped["Anteil Gesamt (%)"] = grouped['Anzahl'] / total_entrytypes * 100 if total_entrytypes else 0
 
     # Mapping Eintragstyp zu Typgruppe
     eintragstyp_gruppen = {
@@ -1643,8 +1302,7 @@ def visualize_language_entrytypes(bib_database):
         barmode="group",
         title=f'Verteilung der Eintragstypen pro Sprache (n={len(df)}, Stand: {current_date})',
         text='Anzahl',
-        labels={'Sprache': 'Sprache', 'Eintragstyp': 'Eintragstyp', 'Anzahl': 'Anzahl', 'Typgruppe': 'Typgruppe'},
+        labels={'Sprache': 'Sprache', 'Eintragstyp': 'Eintragstyp', 'Anzahl': 'Anzahl', 'Typgruppe': 'Typgruppe'}
-        custom_data=['Eintragstyp', 'Typgruppe', 'Anteil Gesamt (%)', 'Anteil innerhalb Sprache (%)']
     )
     layout = get_standard_layout(
         title=fig.layout.title.text,
@@ -1658,19 +1316,9 @@ def visualize_language_entrytypes(bib_database):
     # Ergänzung: Y-Achse logarithmisch skalieren
     layout["yaxis_type"] = "log"
     fig.update_layout(**layout)
-    fig.update_traces(
-        hovertemplate=(
-            "<b>%{x}</b><br>"
-            "Eintragstyp: %{customdata[0]}<br>"
-            "Typgruppe: %{customdata[1]}<br>"
-            "Anzahl: %{y}<br>"
-            "Anteil gesamt: %{customdata[2]:.2f}%<br>"
-            "Anteil innerhalb Sprache: %{customdata[3]:.2f}%<extra></extra>"
-        )
-    )
     fig.show(config={"responsive": True})
     print(tabulate(grouped.sort_values(["Sprache", "Eintragstyp"]), headers=["Sprache", "Eintragstyp", "Anzahl", "Anteil innerhalb Sprache (%)", "Typgruppe"], tablefmt="grid", showindex=False))
-    export_figure_local(fig, "visualize_language_entrytypes", export_fig_visualize_languages)
+    export_figure_local(fig, "visualize_language_entrytypes", export_fig_visualize_languages, bib_filename)
 
 #############
 
@@ -1713,11 +1361,9 @@ visualize_tags(bib_database)
 visualize_index(bib_database)
 visualize_research_questions(bib_database)
 visualize_categories(bib_database)
-visualize_relevance_vs_research_questions(bib_database)
-visualize_relevance_vs_categories(bib_database)
-visualize_relevance_vs_search_terms(bib_database)
 visualize_time_series(bib_database)
 visualize_top_authors(bib_database)
+visualize_top_publications(bib_database)
 data = prepare_path_data(bib_database)
 create_path_diagram(data)
 create_sankey_diagram(bib_database)

deskriptive-literaturauswahl.py

@@ -218,13 +218,14 @@ layout["legend"] = dict(
     itemdoubleclick="toggle"
 )
 layout["yaxis3"] = dict(
-    title=dict(text="Abweichung (ΔSCₙ)", font=dict(color=colors["text"])),
+    title="Abweichung (ΔSCₙ)",
     overlaying="y",
     side="right",
     showgrid=False,
     zeroline=True,
     zerolinewidth=2,
     zerolinecolor='grey',
+    titlefont=dict(color=colors["text"]),
     tickfont=dict(color=colors["text"]),
     anchor="free",
     position=1.0